bttjv5n1

Bechtel Telecommunications

Technical JournalJanuary 2007

Bechtel Telecommunications Technical Journal

Volume 5, Number 1

ADVISORY BOARDJake MacLeod, Principal Vice President and

Chief Technology OfficerS. Rasoul Safavian, PhD, Vice President of Technology,

Americas Regional Business UnitBrian Coombe, Program Manager,

Strategic Infrastructure Group

EDITORIAL BOARDS. Rasoul Safavian, PhD, Editor-in-ChiefRichard Peters, Senior Technical EditorTeresa Baines, Senior Technical EditorBarbara Oldroyd, Technical EditorRuthanne Evans, Technical Editor

GRAPHICS/DESIGNKeith Schools, Art DirectorDavid Williams, Graphic DesignerLuke Williams, Graphic DesignerSusan Gann, Graphic DesignerDiane Cole, Desktop Publishing

TRADEMARK ACKNOWLEDGMENTSAll product or service names mentioned in this BechtelTelecommunications Technical Journal are trademarks of theirrespective companies. Specifically:

AirTegrity Wireless and WiMAX-in-a-Box are trademarks of AirTegrity Wireless, Inc.

Alcatel-Lucent is a trademark of Alcatel-Lucent.Amperion is a trademark of Amperion, Inc.cdma2000 is a registered trademark and certification mark of

the Telecommunications Industry Association (TIA-USA).Centrino, Intel, and Xircom are registered trademarks of Intel Corporation

or its subsidiaries in the United States and other countries.CORBA is a registered trademark of Object Management Group, Inc.,

in the United States and/or other countries.CURRENT Technologies is a registered trademark of CURRENT

Communications Group, LLC.EarthLink is a registered trademark of EarthLink, Inc. Ericsson is the trademark or registered trademark of

Telefonaktiebolaget LM Ericsson. Google Earth is a trademark of Google Inc.HomePlug is a registered trademark of the HomePlug

Powerline Alliance. Huawei is a trademark of Huawei Technologies Co., Ltd.Java is a trademark of Sun Microsystems, Inc.,

in the United States and other countries.Microsoft and PowerPoint are registered trademarks and

Virtual Earth is a trademark of Microsoft Corporation in the United States and other countries.

Motorola is registered in the U.S. Patent and Trademark Office by Motorola, Inc.

Nortel is a trademark of Nortel Networks Limited.P1675 is a trademark of the IEEE. Power Vision is a service mark and Sprint is a trademark of Sprint Nextel.QChat is a trademark of QUALCOMM Incorporated.QUALCOMM is a registered trademark of QUALCOMM Incorporated.Samsung is a trademark of Samsung in the United States or other

countries.Stratellite is a trademark of Sanswire Networks.Tellabs is a registered trademark of Tellabs Operations, Inc..T-Mobile is a federally registered trademark of Deutsche Telekom AG. The Verizon name and Verizon Wireless are registered trademarks of

Verizon Trademark Services LLC or its affiliates in the United Statesand/or other countries.

W3C is a registered trademark (registered in numerous countries) of theWorld Wide Web Consortium; marks of W3C are registered and held by its host institutions MIT, ERCIM, and Keio.

Wi-Fi and Wi-Fi Alliance are registered trademarks and Wi-Fi CERTIFIED is a trademark of the Wi-Fi Alliance.

WiMAX, WiMAX Forum, and WiMAX Forum Certified are trademarks of the WiMAX Forum.

Yahoo! is a registered trademark of Yahoo! Inc. ZTE is a registered trademark of ZTE Corporation.

Contents

Foreword v

Editorial vii

WiMAX™ IEEE 802.16e Plugfest and 1Network Interoperability Testing: Overview and Path ForwardEsmael Dinan, PhD (Bechtel), and Ed Agis, Asha R. Keddy, and Jeremy Rover (Intel)

GIS for Telecommunications 11Paul A. Lukas

Broadband over Power Lines (BPL) 19Lee Lushbaugh and S. Rasoul Safavian, PhD

Service-Oriented Architecture 39Brian Coombe

Technical Aspects of Localization in 47Indoor Wireless Networks (Invited Paper)Muzaffer Kanaan1,2, Mohammad Heidari2, Ferit Ozan Akgül2, and Prof. Kaveh Pahlavan, PhD2

(1Verizon Laboratories; 2CWINS, Worcester Polytechnic Institute)

Fieldable Digital Coherent Interferometric Communication 59and Sensing Application Domains (Invited Paper)Isaac Shpantzer, PhD (CeLight, Inc.)

Solar Energy in Telecommunications 65Glenn A. Torshizi and Mansour Niknam

cdma2000® Wireless Local Loop Evolution and Performance 75Nathan Youell

© 2007 Bechtel Corporation. All rights reserved.Bechtel Telecommunications is a business unit of the Bechtel group of companies, including Bechtel Corporation inthe United States.

Bechtel welcomes inquiries concerning the BTTJ. For further information or for permission to reproduce any paperincluded in this publication in whole or in part, please contact Bechtel Telecommunications, 5275 Westview Drive,Frederick, MD 21703, telephone 301-228-7500 or toll-free 800-946-3232, and ask for the Chief Technology Officer.

Although reasonable efforts have been made to check the papers included in the BTTJ, this publication should notbe interpreted as a representation or warranty by Bechtel Corporation of the accuracy of the information containedin any paper, and readers should not rely on any paper for any particular application of any technology withoutprofessional consultation as to the circumstances of that application. Similarly, the authors and Bechtel Corporationdisclaim any intent to endorse or disparage any particular vendors of any technology.

January 2007 • Volume 5, Number 1 iii

TELECOMMUNICATIONS

The BTTJ is also available on the Web at www.bechteltelecoms.com/jsp/labs/pubs.jsp

It is with great pleasure that I announce that this will be the last edition of the Bechtel Telecommunications Technical Journal (BTTJ). No, we are not terminating the journal, but simply renaming it to coincide with therenaming of Bechtel Telecommunications (a subsidiary of Bechtel Group, Inc.) to Bechtel Communications, Inc.

The new name reflects Bechtel's much broader scope of involvement in the communications arena. Beginning with the next issue, the BTTJ will be renamed the Bechtel Communications Technology Journal (BCTJ) to moreappropriately reflect its mission and objectives. Bechtel's technology activities continue to encompass the design and implementation of traditional telecommunications; however, our current activities also include the design andimplementation of networks ranging from voice telecommunications to video broadcast to satellite to data tobroadband over power lines, and more.

The BCTJ will continue to provide a forum in which to address current technology issues of concern to the globalcommunications community; particularly, operational and network performance issues. The BCTJ will likewisecontinue to provide a platform on which to introduce future technologies before deployment and to attempt toidentify functional and operational benefits and deficiencies. Bechtel Communications is vendor neutral andtechnology agnostic, which positions us to provide an unbiased perspective of the technologies under evaluation, suchas broadband over power lines, service-oriented architecture, and coherent interferometric communications systems,all addressed in this current issue.

This issue begins with a summary of the 5th WiMAX™ 802.16e Global Interoperability Plugfest, which the BechtelCommunications Laboratories (BCL) hosted in our facilities in Frederick, Maryland, in September 2006. This represents the first and only time the WiMAX IOT will be conducted in North America. Previous Plugfests wereheld in Beijing, China; Málaga, Spain; and Sophia Antipolis, France. The Plugfest provided equipment vendors an opportunity to test their latest 802.16e equipment for interoperability with other vendors’ equipment in a secure,neutral environment.

Again, it is with great pleasure that we bring you the latest issue of the BTTJ. We encourage and welcome your feedback regarding the journal. You can submit comments and suggestions for improvement by accessing theBechtel Telecoms Web page at www.bechteltelecoms.com and click on the “Contact Us” section. Past issues of theBTTJ can be downloaded from our Technology Web page at www.bechteltelecoms.com/jsp/labs/pubs.jsp.

May your efforts be safe, productive, and prosperous.

Sincerely,

Jake MacLeodPrincipal Vice President and Chief Technology OfficerBechtel Communications

January 2007 • Volume 5, Number 1 v

Foreword

Editorial

January 2007 • Volume 5, Number 1 vii

Change—in technology and otherwise—is inevitable. In fact, it has been aptly said that the one thing constant in life is change! Perhaps nowhere is change more rapid and profound today than in the technology arena.Particularly in communications and information technology, extraordinary changes are occurring at a

record-breaking pace and scale. The prevalent example is mobile communications. From its analog introduction in the early 1980s to its initial penetration rate of 10 million subscribers by the mid-1990s to its milestone penetration rateof 1 billion subscribers by 2002 to its explosive growth to over 2 billion subscribers by 2006, its direct impact on almost every aspect of daily life is truly astonishing! With all these rapid changes, it has become essential tocontinuously ask: Where are we and where do we go from here? What technologies will prevail, and who will beserving the next 2 billion subscribers?

At the BTTJ, one of our main missions is to provide insight into new trends and directions in communications andinformation technology by examining the technical, operational, and regulatory issues surrounding existing and newdevelopments. The current issue opens with WiMAX™ IEEE 802.16e Plugfest and Network Interoperability Testing:Overview and Path Forward. In this paper, Dinan, Agis, Keddy, and Rover highlight WiMAX interoperability activities, including Plugfest and NWIOT, and look into a new open platform for NWIOT that can help equipment manufacturers and service operators perform tests efficiently and cost-effectively. In the second paper, GIS for Telecommunications, Lukas describes how GIS can be used effectively in communications network planning and design as well as in outside plant operations and workflow optimization.

Next, Lushbaugh and Safavian examine the state of the emerging Broadband over Power Lines (BPL) technology. As athird broadband avenue into homes and businesses (the other two being DSL and cable), electric power lines couldpotentially provide broadband service to the 4 billion people on our planet who currently have access to them! In the paper that follows, Coombe looks into Service-Oriented Architecture and its benefits to carriers in managingnetworks and customers and deploying new services and applications.

We are also pleased to present two exciting invited papers. In Technical Aspects of Localization in Indoor Wireless Networks, Kanaan, Heidari, Akgül, and Pahlavan examine challenges facing accurate indoor localization andlook into several estimation techniques and their associated qualities of estimators. In Fieldable Digital CoherentInterferometric Communication and Sensing Application Domains, Shpantzer proposes a new paradigm based on a layeredarchitecture that unifies optical coherent communications and interferometric sensing. The proposed approach isbased on a digitally stabilized quadrature modulator and a homodyne receiver, augmented by digital noise reduction and channel compensation algorithms.

In Solar Energy in Telecommunications, Torshizi and Niknam show how solar-energy-based systems can provide robust, cost-effective alternative energy solutions to meet communications needs. Rounding out this issue is cdma2000® Wireless Local Loop Evolution and Performance, in which Youell highlights how modern WLL network performance could match that of DSL networks and how WLL can provide new means of high-quality fixed and mobile broadband wireless access.

In closing, I hope you find this new edition of the BTTJ informative and useful. As always, I look forward to yourcomments and contributions. I would also like to take this opportunity to wish you—our readers—a very happy,prosperous, and safe new year!

Happy Reading!

Dr. S. Rasoul SafavianVice President and Editor-in-ChiefBechtel Communications

© 2007 Bechtel Corporation. All rights reserved. 1

INTRODUCTION

Worldwide Interoperability for MicrowaveAccess (WiMAX™) enables the delivery of

“last-mile” wireless broadband access as analternative to cable and digital subscriber line(DSL). The technology behind WiMAX is basedon the Institute of Electrical and ElectronicsEngineers (IEEE) standards. Mobile WiMAXtechnology enables mobility features andattributes for the end user. WiMAX technologycan provide fixed, nomadic, portable, and mobilewireless broadband connectivity without theneed for a direct line of sight with a base station(BS). In a typical cell radius deployment of 3 to 10 km, WiMAX Forum Certified™ systems can be expected to deliver capacity of up to 40 Mbps per channel for fixed and portable accessapplications. This is enough bandwidth tosimultaneously support hundreds of businesseswith T-1-speed connectivity and thousands ofresidences with DSL-speed connectivity. Mobilenetwork deployments are expected to provide up to 15 Mbps of capacity within a typical cellradius deployment of up to 3 km [1, 2]. WiMAX-capable chipsets incorporated into notebookcomputers, ultra mobile personal computers(PCs), personal digital assistants (PDAs), andhandsets will enable portable outdoor broadbandwireless access for the private and public sectors.

One of the main characteristics of WiMAXtechnology is the interoperability of equipmentcertified by the WiMAX Forum™. Certificationhelps boost equipment sales volumes and givesservice providers the flexibility to buy equipmentfrom more than one company with theknowledge that everything will interoperate. TheWiMAX Forum has brought together leaders inthe communications and computing industries todrive the development of a common platform for the global deployment of Internet Protocol(IP)-based broadband wireless services.

IEEE Standard 802.16e-2005 [3], which pertains tothe air interface for fixed broadband wirelessaccess systems, contains options for a number ofphysical (PHY) layers for different frequencybands and region-by-region frequency regulatoryrules. To achieve interoperability, the WiMAXForum has undertaken the development ofsystem profiles specifying which options to use,testing specifications to verify these specificprofiles, and certification laboratories wherevendors can prove that their equipment meetsthese profiles and interoperates.

The WiMAX Forum Certification Working Group(CWG) handles the operational aspects of theWiMAX Forum certification program. The CWGalso plans Plugfests (and prepares their test

WiMAX™ IEEE 802.16e PLUGFEST AND NETWORK INTEROPERABILITY TESTING: OVERVIEW AND PATH FORWARD

Abstract—A primary benefit of WiMAX™ technology is the interoperability of WiMAX-capable equipment,resulting in lower equipment cost and the ability of service providers to purchase equipment from multiplevendors. A WiMAX Plugfest brings together leading equipment vendors from all continents to drive thedevelopment of a common platform for the WiMAX physical and medium access layers. The WiMAX Forum™is also launching a network interoperability testing (NWIOT) program. The goal of NWIOT is to accomplishend-to-end network-level interoperability across WiMAX-capable network components. The architecture isdefined in the WiMAX Forum Network Working Group’s specification. This paper presents an overview ofWiMAX interoperability activities, including Plugfest and NWIOT. The path forward to an interoperableWiMAX-capable mobile technology is described. An open platform is proposed for NWIOT that helpsequipment manufacturers and service operators perform tests efficiently and cost-effectively.

Key Words—architecture, conformance testing, interoperability, network performance, network testing,WiMAX

Issue Date: January 2007

Esmael Dinan, [email protected]

Ed Agis (Intel)

[email protected]

Asha R. Keddy (Intel)

[email protected]

Jeremy Rover (Intel)

[email protected]

plans) at various test laboratories around theworld about every 4 months.

The WiMAX Forum’s network interoperabilitytesting (NWIOT) Task Group (TG) is made up ofrepresentatives from the WiMAX Forum’s serviceprovider, network, marketing, and certificationworking groups. The goal of the NetworkWorking Group (NWG) is to create the networklayer specifications for mobile WiMAX systems,beyond what is defined in the scope of IEEE802.16 (medium access control [MAC] and PHY).

The NWIOT TG is responsible for developing testspecifications for end-to-end network levelinteroperability across all normative referencepoints as applied across the network profilesdefined in the NWG’s specifications. The CWG isresponsible for the execution aspects, similar tothe role it plays in certification.

This paper presents an overview of WiMAXForum Plugfest and NWIOT initiatives andproposes an open platform on which to performtests.

WiMAX FORUM PLUGFEST

Agroup test—commonly called a plugfest—isone of several venues used by numerous

technology consortiums. A plugfest is one means of providing vendors an opportunity to address potential ambiguities and to improve testing scenarios and capabilities for atechnology standard.

WiMAX Forum Plugfests are typically week-longevents carried out at WiMAX Forum-contractedtesting sites primarily to validate and verify theinteroperability of equipment among vendors [4].A vendor is considered to be interoperable once it

Bechtel Telecommunications Technical Journal 2

ABBREVIATIONS, ACRONYMS, AND TERMS

AAA authentication, authorization, and accounting

ASN access service network

ASP application service provider

BE best effort

BRAN broadband radio access network

BS base station

CSN core service network

CWG Certification Working Group (WiMAX Forum)

DSL digital subscriber line

ETSI European TelecommunicationsStandards Institute

GW gateway

IEEE Institute of Electrical and Electronics Engineers

IMS IP multimedia subsystem

IP Internet Protocol

MAC medium access control

MS mobile station

NAP network access point

NCT network conformance testing

NRM Network Reference Model (WiMAX Forum)

NSP network service point

A plugfest is one means of

providing vendorsan opportunity

to address potential

ambiguities and to improve

testing scenariosand capabilities for a technology

standard.

NWG Network Working Group (WiMAX Forum)

NWIOT network interoperability testing

PC personal computer

PDA personal digital assistant

PHY physical

PTCC Protocol and Testing Competence Centre (ETSI)

QoS quality of service

RF radio frequency

SPWG Service Provider Working Group (WiMAX Forum)

SS subscriber station

SUT system under test

TDD time division duplex

TG Task Group (WiMAX Forum)

TWG Technical Working Group (WiMAX Forum)

VSA vector signal analyzer

WiMAX™ Worldwide Interoperability for Microwave Access (Although synonymous with the IEEE 802.16 standards suiteand standardized by IEEE, WiMAX is a certification mark promoted by the WiMAX Forum.)

January 2007 • Volume 5, Number 1 3

has demonstrated that its hardware can send andreceive packets of data with two other vendorsinvolving BSs and subscriber stations (SSs) for aselected certification profile. Before a Plugfestvenue occurs, participating vendors agree on a set of radio frequency (RF)/PHY characteristicswithin a given certification profile. In allinstances, a minimum of three vendors must beavailable to conduct a suite of selectedinteroperability testing scenarios within a given certification profile. The WiMAX Forumrequires a minimum of five or six vendors for any Plugfest.

The key objectives of a WiMAX Forum Plugfestare to:

• Identify where there may be differingstandards interpretations that must beresolved

• Identify interoperability problems that maybe firmware or software related

• Encourage open and unambiguous technicaldiscussions of the test scenarios and thestandard, with a means to refine them

• Prepare a vendor to submit its products forformal certification testing

• Continuously improve the quality ofinteroperability testing to ensure a viableWiMAX certification process

• Make improvements for implementingfuture group testing venues

The WiMAX Forum Plugfest held fromSeptember 24 to October 1, 2006, was hosted by Bechtel Telecommunications (Frederick,Maryland, USA) in collaboration with Centro deTecnología de las Comunicaciones (Málaga,Spain)—newly renamed AT4 wireless—and theWiMAX Forum. This was the fifth Plugfest that the CWG conducted where service flows for quality of service (QoS) traffic, network entry procedures, and implementation on theclassification of data packets for mobile WiMAXwere all achieved, as part of preparing WiMAXForum certification testing for Wave 1 of mobileWiMAX. While the focus of this venue waspredominantly mobile WiMAX, several vendorsalso demonstrated interoperability testing fornomadicity and portability.

The event was organized as a cooperativeagreement between Bechtel Telecommunicationsand the WiMAX Forum. Bechtel providedlogistical and engineering services as the primaryhost for this first WiMAX Forum Mobile Plugfest.

This event was also important as the first WiMAXForum Plugfest in North America and the firstpublic WiMAX Forum Mobile Plugfestworldwide. Interoperability scenarios andprofiles for this venue were developed jointly bythe CWG and the WiMAX Forum TechnicalWorking Group (TWG) for mobile WiMAXdevices. The interoperability testing scenariosincorporated the IEEE 802.16 and EuropeanTelecommunications Standards Institute (ETSI)HiperMAN standards in the joint development ofprotocol conformance test specifications that areone of the essential elements of the WiMAXcertification program. The broadband radioaccess network (BRAN) Technical Committee isthe home to these activities at ETSI, whereHiperAccess and HiperMAN standards andWiMAX/HiperMAN test specifications aredeveloped with the extensive support of ETSI’sProtocol and Testing Competence Centre (PTCC).Most of the work lends itself to improvinginteroperability among vendors. This FirstMobile Plugfest event greatly contributed tobringing equipment vendors closer to achievingWiMAX certification, which is important tomanufacturers, telecommunications carriers, andservice providers alike.

PLUGFEST CONFIGURATIONS AND TEST PLANS

Five system test configurations were definedand used at the Mobile Plugfest. A system

under test (SUT) is defined as a networkconsisting of one BS and one to three mobilestations (MSs). The system includes, whennecessary, monitoring devices such as a WiMAXprotocol analyzer and/or a vector signal analyzer(VSA). The following specific configurations wereused at the Plugfest:

• SUT1: Single BS + Single MS—one vendor

• SUT2: Single BS + Single MS—two vendors

• SUT3: Single BS + Two MSs (same vendor)

• SUT4: Single BS + Two MSs (differentvendors)

• SUT5: Single BS + Three MSs (threedifferent vendors)

Devices in each SUT were interconnected bywired means. QoS testing was conducted in termsof interoperability to emulate the real final userexperience (i.e., transmitting data according tothe QoS parameters defined and checking that theQoS of a service flow was not affected by otherbest-effort [BE] data transmissions).

An SUTis defined as

a network consisting of one BS and

one to three MSs.


As an example, Figure 1 shows SUT5. In this testconfiguration, a single BS is connected to threeMSs. Each MS could be from a different vendor,or two MSs could be from the same vendor, orone MS could be from the BS vendor, dependingon the testing schedule at the Plugfest.

Certification profiles tested at the Mobile Plugfestconsisted of:

• 2.3–2.4 GHz, 5/8.75/10 MHz, time division duplex (TDD)

• 2.496–2.69 GHz, 5/10 MHz, TDD

• 3.4–3.6 GHz, 5/7 MHz, TDD

• 4.935–4.990 GHz, 5 MHz, TDD

Test PlanTest scenarios for the Mobile Plugfest wereorganized into the following three groups,according to the expected flow described below,to achieve data packet exchange among vendorhardware:

1. Network Entry Procedure

2. Traffic Connections Establishment

3. User Data Transfer

The scenarios were sequenced as shown in Table 1, according to complexity and feasibility.

Test scenarios were developed to test eachfunctionality. To consider all importantfunctionalities needed to achieve interoperabilityamong different vendor devices, five testscenarios were defined for the Plugfest. Becausethey are not designed as conformance tests, thescenarios do not determine if a product conformsto the standard. Rather, they provide a method ofisolating and resolving problems within WiMAX-capable devices that may affect their ability to interoperate [5].

Path ForwardA total of 19 vendors and 3 test equipmentvendors participated in the WiMAX MobilePlugfest held at Bechtel Telecommunications inFrederick. This turnout represents the largestPlugfest in the history of the WiMAX Forum and reaffirms one of the key elements of the WiMAX Forum’s charter: To develop aframework for high performance end-to-end IP network architecture supporting stationary,portable, and mobile usage models.

This Plugfest continued the evolution of WiMAXtesting methodology and tools, such as WiMAXdevice protocols and MAC and PHY uplink anddownlink monitoring. Ensuring optimal testinghas resolved connectivity problems amongdifferent vendors. There are many opportunitiesto improve the test processes and scenarios. The test scopes at the Mobile Plugfest covered aselect number of IEEE 802.16e PHY and MAClayer features. In upcoming Plugfests, BS handoff scenarios and mobility could be verified,as well as other QoS features. The testconfiguration shown in Figure 2 is proposed forfuture Plugfests to verify more BS features andtest scenarios.

1. Network Entry Procedure 1.1 MS(s) Synchronize to BS1.2 Ranging1.3 Capabilities Negotiation1.4 Authentication (Not Used)1.5 Registration

2. Traffic Connections Establishment

2.1 Service Flow Provisioning2.2 Service Flow Activation

3. User Data Transfer 3.1 Downlink Ping3.2 Uplink Ping

Table 1. Test Scenario Structure for WiMAX Mobile Plugfest

Figure 2. Proposed Test Configuration Covering Handoff Scenarios

A total of 19 vendors and

3 test equipmentvendors

participated in the WiMAX

Mobile Plugfest held at Bechtel

Telecommunicationsin Frederick.

Figure 1. SUT5 Configuration (Equipment Made by Different Manufacturers)

MS1

MS2

BS

MS3

Handoff Mechanism

BS1

BS2

MS3

MS4

MS1

MS2

R8

R1

R1

R1

R1


WiMAX NETWORK INTEROPERABILITY

Plugfests and the certification process provideinteroperability testing and certification

testing for the PHY and MAC layers. CurrentWiMAX certification is not enough for manywireless operators.

To achieve full end-to-end interoperability, the higher layers in the stack and networkinteroperability need to be considered.

Carriers seek to provide high-performance, high-reliability products and need assurance that allsystem nodes are operating as designed. Theywant to explore the technical feasibility ofnetwork-level interoperability, including end-to-end network testing scenarios. To achieveNWIOT, two things need to happen:

1. There must be interoperability betweendifferent network elements across vendorimplementations, also known as infra-structure interoperability. This means thatthe infrastructure elements in the networkcan interoperate. An example of infra-structure interoperability is an access servicenetwork from one vendor being able to handoff to an access service network from anothervendor, as illustrated in Figure 3.

2. There must be interoperability of userdevices across WiMAX network implemen-tations, also known as network conform-ance testing (NCT). This means that any

WiMAX-capable device can connect to anynetwork. (See Figure 4.)

NWIOT benefits include:

• For operators—A means to help witheconomy of scale while reducing capital andoperational expenses. NWIOT also helpsfacilitate roaming.

• For vendors—More market opportunitiesand economies of scale with WiMAX. (On acautionary note, the WiMAX Forum has to be careful to provide room for innovationand differentiation.)

• For customers—Improved global roamingcapabilities and reduced cost of service.

As previously stated, the NWIOT programcrosses WiMAX Forum working groups,involving the NWG, the Service Provider

Figure 3. Infrastructure Interoperability

BSFunctions

Multi-sectorCell Site

IPNetwork

Standalone CSN2 or Incumbent IP Core

Gateway

MSS1

SS2

ASN 3

ASN 1

BS + ASN GWFunctions

Some BS +All ASN GWFunctions

Standalone CSN1 or Incumbent IP Core

Gateway

Single/Multi-sectorCell Site

Ingress Router/Gateway

Egress Router/Firewall

AAA(Radius/Diameter)

Multi-sectorCell Site

Private IMSNetwork

Ingress Router/Gateway

Egress Router/Firewall

AAA(Radius/Diameter)

Multiple ASNVendor

Implementations

ASN GWFunctions

ASN 2

Multiple Usage Scenarios(phased)

Private IMSNetwork

Some BSFunctions

IPNetwork

Multiple Operators

Multiple Interfaces

Figure 4. Interoperability of User Devices

WiMAXNetwork B

WiMAXNetwork A User Device

Testing

Carriers want to explore the

technical feasibilityof network-levelinteroperability,

including end-to-end networktesting scenarios.


Working Group (SPWG), and the CWG. From atechnical standpoint, test specifications aredeveloped by a NWIOT technical specificationTG under the NWG. The technical coordinator ofthe NWIOT program coordination committeeperforms the necessary NWG/SPWG/CWGtechnical coordination, and the CWG handles theexecution aspects of NWIOT.

The NWIOT TG’s charter is twofold (reflectingthe two aspects of NWIOT):

1. Developing test specifications for end-to-end, network-level interoperability across all normative reference points as applied across network profiles defined in theNWG’s specifications. This ensures thesuccessful interworking of WiMAX networkelements while meeting applicable Stage 1requirements.

2. Developing test requirements for MSs/SSsbased on WiMAX Forum NWG releases. The test requirements are used by the TWGand other relevant entities to facilitate con-formance testing of the specified features.NWIOT provides the necessary guidanceand clarifications among the external andinternal working groups that are responsiblefor test development and execution.

The NWIOT TG establishes strong liaisons withother working groups to establish a successfulprogram. It works with the SPWG to clarifyexisting requirements that are targeted by theNWG and to provide usage scenarios. The TGworks with the CWG to provide necessarytechnical support throughout NWIOT detailedtest plan development and execution (forexample, selection of a test house and a test scriptdeveloper). The NWIOT TG works within theNWG to align the NWIOT test specifications with the Stage 2/Stage 3 specifications and with the NWG’s feature list and to review the NWG specifications.

ROLE OF PERFORMANCE

Carriers need to provide high-performance,high-reliability products and need assurance

that all system nodes are operating as designed.Fierce QoS competition with other serviceproviders is the main concern. Currently,network interoperability alone does not test forthis in the WiMAX Forum.

Equipment quality can affect networkperformance and customer satisfaction and,therefore, produce customer churn. Carriers

want to test network performance for threeprimary reasons:

1. Customer Care—Carriers/service providersown the relationship with the end user.Customer care expenses can increase thesubsidy of new devices, calls into customercare, and foot traffic into retail stores.Ultimately, reduced customer satisfactioncan affect customer churn. When scaled up to support carriers with millions ofcustomers, customer care expenses can be anenormous burden.

2. Network Maintenance—Network mainte-nance costs increase with the use of faultyequipment and the resulting increasednumber of alarms.

3. Network Resources Protection—Ultimately,network resource protection typicallytranslates into air interface capacity (in licensed bands, spectrum capacity is a precious resource). Poor transmitterperformance can induce unwantedinterference on the target or adjacentchannels and reduce air interfacethroughput and capacity. Poor MS receiverperformance can cause BSs to allocate moreenergy per bit to that user to compensateand cause an increase in retransmitted data.Ironically, some poor receiver designs can unintentionally emit interference ontothe channel.

NWIOT TEST SETUP

Given that the goal of NWIOT is pioneering in nature, there are technical, logistical,

business, and operational questions regarding its achievement. Test case examples includeauthentication, accounting, and end-to-end callset-up. From the WiMAX Forum’s perspective,the focus is on the technical specifications, with the logistics and operational aspects still to be determined.

Ongoing high-level questions to be agreed on forNWIOT scope include:

• How much network interoperabilitygranularity is required?

• What is the availability timeline?

Figure 5 illustrates the NWG’s NormativeReference Model (NRM), consisting of thefollowing logical entities: MS, access servicenetwork (ASN), and core service network (CSN)and clearly identified reference points for

Equipment qualitycan affect network

performance and customer

satisfaction and,therefore, produce

customer churn.


interconnecting these logical entities. The figuredepicts key normative reference points R1–R5.Each logical entity—MS, ASN, and CSN—represents a group of functions. Each functionmay be realized in a single physical device or may be distributed over multiple physicaldevices. The grouping and distribution offunctions into physical devices within a logicalentity (such as an ASN) is an implementationchoice; a manufacturer may choose any physical implementation of functions, eitherindividually or in combination, as long as the implementation meets functional andinteroperability requirements.

The intent of the NRM is to allow multipleimplementation options for a given logical entitywhile achieving interoperability among differentrealizations of these entities. Interoperability isbased on the definition of communicationprotocols and data plane treatment betweenlogical entities to achieve an overall end-to-endfunction, for example, security or mobilitymanagement. Thus, the logical entities on eitherside of a reference point represent a collection ofcontrol and bearer plane end-points.

The ASN defines a logical boundary andrepresents a convenient way to describe anaggregation of functions and the correspondingmessage flows associated with access services.The ASN represents a boundary for functionalinteroperability with WiMAX clients; the WiMAX connectivity service functions as anaggregation of functions embodied by differentvendors. CSN is defined as a set of networkfunctions that provide IP connectivity services to the WiMAX subscriber(s). A CSN maycomprise network elements such as routers;authentication, authorization, and accounting(AAA) proxy/servers; user databases; andinterworking gateway (GW) devices.

An ASN has three implementation options, called profiles by the NWG. In Profiles A and C,the BS-ASN GW interface is exposed, whereas inProfile B, implementation is a “black box.” Toaccommodate these different implementationsand to phase the NWIOT, there are two levels ofinfrastructure interoperability:

• Level 1: ASN-profile-independent operability

• Level 2: ASN-profile-dependent operability

Figure 6 shows a sample end-to-end profile forprofile-independent interoperability tests. This level:

• Assumes that the ASN is a logical entity (i.e., a “black box”)

• Concentrates on tests with multiple usagesmodels, multiple operators, and multipleASN implementations

• Affects the R1, R2 (not shown), R3, and R4interfaces

Figure 5. NWG Normative Reference Model [6]

SS/MS ASN

NAP

R1 R3 R5

R4

R2

R2

AnotherASN

ASP Networkor Internet

Visited NSP

ASP Networkor Internet

Home NSP

CSN CSN

Figure 6. End-to-End Profile-Independent Network Interoperability

R1 R3ASN

Gateway

R4

Base Station

Base Station and ASN Gateway Functions


Figure 7 shows a sample end-to-end profile forprofile-dependent interoperability tests. This level:

• Tests interoperability between differentvendors within an ASN

• Requires vendors from the same profile(hence the term profile dependent)

• Affects the R1, R2 (not shown), R3, R4, andR6 (internal ASN) interfaces

The NWG needs to consider all options and theirfeasibility and timeline issues. Both ASN-profile-independent and -dependent tests can bedeveloped simultaneously and be available.Although these tests address all operatorrequirements, the level of complexity may be toohigh to implement the test setup in one step. Testplans can be defined for mobility, and subsets canbe established for stationary profiles. Test setupstarget full mobility, and tests for otherdeployment models are subsets of the fullmobility tests. Test plans can be developed withcomplete mobility in mind (the most complexsituation) so that there are no forwardcompatibility issues. This will reduce operationaland process overhead because there is only onecomplete test release. Advantages of thelaboratory setup are the controlled, repeatableenvironment and the full automation of tests.

In addition to laboratory testing in a controlledconductance environment, testing can beperformed in a real network environment withover-the-air transmission. Over-the-air testingreflects the real field environment. Figure 8shows an example of WiMAX coverage plots and

an actual cell site photo for an over-the-airNWIOT laboratory. Performance and NWIOTtest cases can be executed in this environment,including end-to-end call setup, handoff,mobility, authentication, accounting, and end-to-end system performance measurements. While afield environment does not provide the samelevel of control as a laboratory environment, anadvantage of the field environment is that itprovides a user-experience aspect to the end-to-end flow.

Testing that would take considerable time duringa maintenance window in a commercial traffic-bearing network can be carried out over a muchshorter time period in such an environment.Further, testing can be accomplished withoutimpact on market network operations or onsubscribers. Manufacturers and service providershave practically unlimited possibilities to testnetwork performance, interoperability, andconformance in such an environment.

CONCLUSIONS

Together, WiMAX certification, Plugfests, andNWIOT constitute a means of providing

interoperable end-to-end WiMAX equipment.The certification process is divided intodeveloping conformance testing and inter-operability testing. The Plugfest is a preview offull interoperability testing that allows vendors toget an early look at how well their equipmentinteroperates. NWIOT ensures multivendorinteroperability of an end-to-end WiMAXnetwork based on NWG specifications. In

Test plans can be developed

with completemobility in mind

(the most complex situation)

so that there are no forward

compatibilityissues.

Figure 7. End-to-End Profile-Dependent Network Interoperability

R1

R3

R4

R8

R6

R6

ASNGateway

Base Station

R1

Base Station


addition to tests performed in the laboratory,over-the-air NWIOT is important because itreflects the real field environment. An over-the-air environment that includes actual WiMAX cellsites provides a common platform for equipmentmanufacturers and service operators to performthe tests efficiently and cost-effectively.Furthermore, the tests can be accomplishedwithout impact on market network operations oron subscribers.

TRADEMARKS

WiMAX, WiMAX Forum, and WiMAX ForumCertified are trademarks of the WiMAX Forum.

REFERENCES

[1] “Mobile WiMAX—Part I: A Technical Overviewand Performance Evaluation,” WiMAX Forumwhite paper, August 2006.

[2] “Mobile WiMAX—Part II: A ComparativeAnalysis,” WiMAX Forum white paper, May 2006.

[3] IEEE Standard 802.16e-2005 — “Amendment toIEEE Standard for Local and Metropolitan AreaNetworks—Part 16: Air Interface for FixedBroadband Wireless Access Systems—Physical and Medium Access Control Layers for Combined Fixed and Mobile Operation in Licensed Bands,” February 2006.

[4] “WiMAX Mobile Plugfest White Paper,” WiMAX Forum, October 2006.

[5] “Guidelines for Scenarios and Profiles for VendorInteroperability Between Mobile WiMAX Devicesin WiMAX Forum™ Mobile Plugfest,” WiMAXForum, September 2006.

[6] “WiMAX End-to-End Network SystemArchitecture,” WiMAX Forum public document,November 17, 2006, revision.

BIOGRAPHIESEsmael Dinan is principalengineer/technologist – BechtelTelecommunications. A keyspecialist for Bechtel’s GlobalTechnology Team, he has been leading the effort tosupport customers with plansfor WiMAX equipment inter-operability and conformancetesting. Dr. Dinan has per-

formed numerous key wireless technology assignmentsand has been instrumental in many aspects of thebusiness unit’s research activities, as well as on theCingular RF engineering project. He designed andengineered an RF engineering data managementsystem, developed Cingular project RF engineeringprocesses and procedures, designed and optimizedUMTS networks, and verified and tested Dupontcryogenic TMA performance.

Before joining Bechtel in 2002, Dr. Dinan was productmanager for the GMPLS control plane of the RAYStarDWDM optical switch at Movaz Networks, and leadnetwork architect at Worldcom. He has conductedresearch and development on access methods andperformance modeling of 3G wireless communicationsand high-speed optical networks.

Dr. Dinan received his PhD in Electrical Engineeringfrom George Mason University, Virginia, and is aregistered Professional Engineer in Maryland. He has authored or co-authored more than 25 conferencepapers and journal articles, including nine papers in previous issues of the Bechtel TelecommunicationsTechnical Journal, and has two patents on novelsignaling mechanisms developed for 3G cellularnetworks. He is a member of the Institute of Electricaland Electronics Engineers.

Figure 8. Example of WiMAX Coverage Plots and Actual Cell Site Photo for Over-the-Air NWIOT Activity

–100 to –90 dBm–90 to –80 dBm –80 to –70 dBm–70 to –60 dBm–60 to –50 dBm–50 to 0 dBm


Ed Agis, a market developmentmanager for Intel® Corpo-ration’s Mobility WirelessStandards and TechnologyGroup, is also the co-chair of theWiMAX Forum’s CertificationWorking Group and a memberof the forum’s Technical andMarketing Working Groups. He is also actively involved in

the IEEE 802.16 standards body and responsible for the development of the certification testinginfrastructure of the WiMAX Forum. Prior to his current position, Ed was in Intel’s Wireless ProductDivision, where he was responsible for marketingprograms strategy and development for their wirelessnetworking planning strategy.

Before joining Intel in January 2001, Ed was director ofmarketing and business unit manager for AccessProducts at Xircom®, where he led the launch ofnumerous mobile access products. Earlier, as the WW product marketing manager for Advance SystemsSolutions and PCI Bus Products at Texas Instruments,Ed led the launch and market development of TI’s PCCard Controller, PCI bridge chips, and Low VoltageLogic chips.

Ed holds a BS degree from the Air Force Academy, Colorado; an MBA in Management from the University of Southern California (Los Angeles)/Amber University (Texas); and an MBA inOperations/Product Marketing from Amber University.

Asha R. Keddy is a seniorengineering manager on theMobile Systems Research team for Intel Corporation’sCorporate Technology Labs. She has more than 10 years of experience in wireless andmobile broadband technology.She drove the effort at theWiMAX Forum to form the

Network Interoperability Task Group and currentlyserves as a vice chair. She is responsible for drivingIntel’s strategic efforts around WiMAX performanceanalysis and characterization, network interoperability,and cross-layer research. Asha has also led Intel’s effortsin the wireless fidelity area, including end-to-endinteroperability of the international roaming accessprotocols; end-to-end performance characterization ofwireless networks; and innovative test technologymethods for Intel Centrino® Mobile Technology. Sheholds multiple patent filings and papers in these areas.

Asha received her BE degree in Computer Engineeringfrom Bombay University, India, and her MS degree in Computer Science from Clemson University, South Carolina.

Jeremy Rover is a seniornetwork engineer on the MobileSystems Research Team at IntelCorporation. His expertiseincludes network performancecharacterization, interoperability,and interference analysis forWiMAX technologies. Withinthe WiMAX Forum, Jeremychairs the Specialized Test

Equipment Sub-group under the Certification WorkingGroup. He has also been actively involved in definingNetwork Interoperability Testing.

As part of Intel Centrino Mobile Technology team,Jeremy worked extensively with wireless fidelitytechnologies and networking. He was the Wi-Fi™Alliance editor for the Managed and Public Access TestPlan. Jeremy has also driven plugfests within the Wi-FiAlliance and at ETSI. He has filed multiple internationalpatents on innovative testing technologies and design.

Jeremy received his BA degree in Computer Sciencefrom Linfield College, Oregon.


INTRODUCTION

Over the past 30 years, geographicinformation systems (GISs) have evolved

from vector-based, computer-aided mappingtools to fully integrated spatial solutionsplatforms. The GIS has become a jack-of-all-trades with capabilities ranging from interactiveWeb-based mapping services to three-dimensional desktop modeling and analysis. Themapping service buzz words these days areGoogle Earth™, Yahoo!® maps, and Microsoft®

Virtual Earth™ [1]. These online mappingapplications have changed the way we live andwork with maps by providing new sources forresearch and data gathering through access tohigh-resolution aerial imagery and point-of-interest data, combined with old-fashionedaddress locators. The fascination with theseflashy applications has left us with areinvigorated interest in discovering what else a GIS can do for us.

The GIS has long been embraced by sciences andindustries as diverse as demographics, medicine,utilities, agriculture, urban planning, biology,advertising, and transportation. The utilitiesindustry, in particular, was an early adopter ofGISs in the form of automated mapping/facilitiesmanagement (AM/FM) applications [2]. Butwhere does telecommunications fit into all of this?

The wireless industry quickly embraced GISs asradio propagation modeling tools, allowingnetwork engineers to rapidly estimate coverage

characteristics before a network was launched.These tools have also enabled engineers to planand optimize changes to existing deployments,letting them view in real time the theoreticalresults of potential changes. Most of the earlynetwork propagation tools were designed arounda GIS platform, linking GIS and networkplanning, sometimes without the user’s knowingthat a GIS was involved or what it was. [3]However, adoption of GISs by thetelecommunications industry as a whole is stillincomplete. We are realizing the benefits ofgeospatial data analysis, but integration could beprogressing faster.

WHAT IS A GEOGRAPHIC INFORMATION SYSTEM?

AGIS is commonly perceived as a singletechnology, usually a software application,

used to create and display cartographicinformation. In practice, however, a GIS consistsof five components: software, data, procedures,hardware, and people. These five componentswork together to capture, store, retrieve, analyze,and display geographically referencedinformation. While computer-aided design(CAD) and mapping applications can displayspatial information, a GIS has the addedcapability to analyze spatial data throughattribute and location analysis or spatialmodeling. Adding a relational database furtherenhances the capability of a GIS to solvecomplicated spatial problems. [4]

GIS FOR TELECOMMUNICATIONS

Abstract—The telecommunications industry is on the verge of a GIS revolution. Using a central corporatedatabase, a GIS can now serve customized data to sales, operations, engineering, customer service, and even thecustomers themselves. What makes this possible is the ability of a GIS to reach into a database and extract spatialinformation targeted to meet a user’s specific needs. This paper focuses on the use of GISs in the networkplanning and design aspect of the telecommunications industry. In particular, the paper highlights the potentialapplications of GIS in outside plant operations and the use of spatialized databases to improve workflow.

Key Words—geographic information systems, GIS, outside plant, spatial database, telecommunications


Paul A. [email protected]

In the telecommunications world, a GIS is ideallysuited for network planning and development.The ability to layer information onto the earth’ssurface, complete with attribute data, allowsengineers the unique ability to model and assessa network from the office. This saves valuabletime and reduces the number of trips, if any, thatthe engineer must make to the field. Furthermore,the powerful automation capabilities offered by a GIS increase the speed and accuracy of thenetwork design process and can help reduce, and even eliminate, the downstream impacts of

design-phase errors on cost and schedule duringthe network deployment phase. Rule-basedfeatures found in a GIS can also offer networkdesigners the ability to produce better products, optimized for cost, shortest routingdistances, or other user-defined metrics. The skill level and design time involved in hand-producing comparable designs would besignificantly higher.

GIS FOR WIRELESS

The GIS is already an essential tool in thewireless industry. Most wireless network

engineers are familiar with the GIS as thebackbone of many wireless design tools alreadyin use. By incorporating digital elevation models(DEMs), land clutter (LC) data, and buildingelevation models, wireless engineers are able toassess radio coverage before the network is built,identify areas that require enhanced capacity orcoverage, and plan for trends in network andapplication performance. However, a GIS can beused even before engineers set pen to paper. Formarketing efforts, population data can be addedto enable network traffic and resource utilizationcalculations. Aggregating block-level populationdata into rough coverage rings yields statistics on expected network complexity and requiredcapital expenditures before network design evenbegins. Such numbers can be extremely effectivewhen used in business development proposalsbecause they originate from real-world modelssuch as the example shown in Figure 1.

A GIS can further be used in businessdevelopment efforts at both the strategic andoperational scales to determine where coverageexpansion should be directed and how theensuing network should be deployed.Demographic data can be leveraged to identifypopulation centers (as, for example, shown in Figure 2) and areas of high income. Ademographic approach removes the need forstrategic planners to “throw darts at the map.”Income estimates further allow engineers totarget geographic areas with high disposableincome that will be most likely to subscribe tonew services. The benefits of this approach areobvious and are further validated by itsincreasing use in diverse business lines. Simplyby picking up a business atlas or doing Internetresearch, planners can identify underservedmarkets and provide opportunities for revenuegeneration. Using a GIS allows the engineer toadd existing and competitive coverage to the mapto improve the context of the data provided.


Figure 1. LU/LC Data Created with the Assistance of Infrared Satellite Imagery


AM automatic mapping

CAD computer-aided design

DEM digital elevation model

FDA fiber distribution area

FDH fiber distribution hub

FM facilities management

GIS geographic information system

GPS global positioning system

LC land clutter

LU land use

NID network interface drop

SDSFIE Spatial Data Standard for Facilities, Infrastructure, and the Environment

SQL structured query language

A GIS is ideallysuited for network

planning anddevelopment.

For example, high speed wireless Internet accesscan be targeted to neighborhoods with apopulation that is inclined to adopt newtechnology. Deploying such services in areas less likely to subscribe wastes resources andslows deployment to areas more likely to desirethe service. Overlaying population, income,educational level, and age datasets enables areasto be identified where residents have greaterdisposable income and are culturally predisposedto purchase wireless services. Adding propensityindexes further allows the engineer to zero in onspecific target communities.

Post-deployment operation can benefit from aGIS as well. After the network is designed andbuilt, the design can be viewed in the GIS foroperational tuning. By creating a detailed map ofa service area, including antenna locations andazimuths, and overlaying the propagationmodels created in the design phase, engineershave a powerful tool for understanding baselineoperations. Drive test data can be loaded into the map to identify areas of real-world service

degradation. By linking base station informationextracted from the drive test data to actualantenna configurations, engineers can easilyidentify and correct poor antenna tuning. “What if” scenarios can be run repeatedly until adesign flaw has been corrected satisfactorily.

Adding base station metrics to the drive test mapfurther helps to identify tuning irregularities.Beyond allowing location information to beviewed on a map, a GIS is capable of runningstructured query language (SQL) queries on data attributes—both geographic and tabular.Augmenting antenna data with base stationidentification information enables the engineer to identify drive test results that lie outsideminimum performance requirements. Theresulting subset of drive test results may then bequeried against the tower database to identifyspecific antennas responsible for the poor signalperformance. This information, viewed on a map,allows the engineer to assess if the antenna ispoorly tuned or if a neighboring antenna is notperforming as intended. Thus, the engineer can


A detailed map of a service area is a powerful tool

that allows the engineer

to focus on solvingthe actual problem,

rather thanspending time on

trial-and-errorsolutions that

may not address the core issue.

Figure 2. Population Density Map


focus on solving the actual problem, rather thanspending time on trial-and-error solutions thatmay not address the core issue. [5]

GIS FOR OUTSIDE PLANT

The role for the GIS in infrastructuremanagement was pioneered by the gas,

electricity, and water utility industries. A GIS isideally suited for outside plant design andmanagement for telecommunications as well.While outside plant was once relegated to a small, specialized community within thetelecommunications world, recent industrychanges have brought a renewed focus to this area. The growth of the Internet, high-definition television, video-on-demand, andother interactive multimedia services has causedend-user demands for bandwidth to skyrocket.Carriers have addressed this requirement bydeploying new outside plant networks, primarily using high-capacity, fiber-optic cable.Deploying new networks often requires revisiting infrastructures placed decades ago;portions of the network could even be more than a century old. [6]

Engineering designs for this infrastructure arelabor intensive due to the physical field survey,

address reconciliation, network dimensioning,and quantity takeoff tasks. Furthermore,incorporating the volume of legacy data toproduce an effective network design mayintroduce a high occurrence of defects. Manuallyimplementing the network design and reviewprocess makes it more difficult to discoverdefects. Correcting these defects often results in acomplete design rework, resulting in cost anddesign cycle time overruns. Because the networkdesign represents an initial stage in the designcycle, it exerts a considerable impact ondownstream items.

Network designers previously relied on CADsystems to support the design efforts; however,GIS advances have positioned them to provide significantly enhanced functionality fordesigning and engineering outside plant systems.The renewed attention to outside plant in thetelecommunications industry coincides well withthe recent GIS developments.

A GIS offers a unique capability over standardCAD applications by allowing disparate datalayers to be assessed independently of theirphysical data attributes. Information such aslength, depth underground, number of cables, orterminal size can be stored as part of the featureinformation (see example in Figure 3). Land-baseinformation such as plat maps can be viewed asindividual raster (pixel) files or vector datasets.Engineers can design network routes directly inthe GIS, registering the data directly over the platmaps, and then export the information to a CADdrawing if necessary. Existing utilities, right-of-way data, and customer information may also beviewed directly in the GIS during the designprocess. The ability to view multiple sources ofinformation at once while designing the networkminimizes the number of field redlines andrevisions necessary before final approval. [7]

Ideally, the GIS becomes a central datamanagement application that combinesinformation from multiple sources. Clientinformation from proprietary legacy databasescan be imported along with land-baseinformation from government sources and thenbe viewed within a single application. The resultis a map that can be printed and taken to the fieldas a single reference source. Equipping fieldcrews with tablet or laptop computers connectedto global positioning system (GPS) receiversmakes the process even more efficient. Fieldcrews are able to view pertinent information onsite, make corrections to attribute data, identifydata irregularities, and relocate incorrectly

The renewedattention to

outside plant in the

telecommunicationsindustry

coincides well with the recent

GIS developments.

Figure 3. Example of Attribute Data that May Be Stored for aFiber-Optics Terminal


positioned infrastructure or add previouslyunknown infrastructure with consistent precisionthat may be relied upon in the office withconfidence. Figure 4 provides an illustration ofthis error-checking capability.

Data from field crews can be downloaded to theinfrastructure database in the office and madeavailable to the design engineers. Incorrect datacan be updated and new features added withoutpassing information back and forth multipletimes, thereby reducing the number of trips to thefield to verify or re-verify data. From this startingpoint, design engineers may begin their workwith confidence that what they see on theirscreens (see example in Figure 5) is an accuratemodel of the real world. The point that accuratedata is used at the outset of the design processcannot be emphasized enough. The time saved bynot having to revise redlines, often multipletimes, results in a more efficient design process.

Automating the design process reduces designturnaround time even further. In a recent projectwhere a GIS was used only as the design tool, the initial network design time was reduced by

two-thirds (not including field checking andredlines resulting from errors beyond the scope ofthe GIS application). For this particular project, acustom GIS application was developed thatallowed the engineer to create customer records,service areas, wiring groups, and cable runs;place terminals; and locate network interfacedrops (NIDs). The benefits of automating theprocess included creating a tabular geospatialdatabase that stored not only the individualgeographic features, but relevant attribute data as well. Included in the attribute data was feature identification information that was usedto create a network topology. [8] The informationin the database could then be incorporateddirectly into automated design summaries (seeexample in Figure 6) that were used for errorchecking, creating a bill of materials, andgenerating a variety of other reports.

After the physical cable design was completed,the underlying network topology was used todetermine the network’s service dimensioning.From the fiber distribution hub (FDH) to the NID,and based on regional requirements, every strandof cable was routed at the click of a button. Asidefrom providing the ability to leverage networktopology, the attribute data from each feature wasused to automatically create material quantitytakeoffs. These two examples show how a GIScan be used to cut design time and allow theengineer to focus on the actual network design,rather than spend time creating reports. Thespeed with which reports can be generated also

Figure 6. Automated Design Summary

In a recent projectwhere a GIS

was used only asthe design tool,

the initial networkdesign time was reduced by two-thirds.

Figure 4. Example of Error-Checking for a Fiber-Optics Terminal Dataset

Figure 5. Raster Plot Map Overlaid with Digital Design Data


contributes to the ability of the engineer toexperiment with “what if” scenarios and compareseveral different designs quickly and withoutmajor rework.

GIS software packages require experienced users.Employing a GIS out of the box may requireadding a cadre of GIS professionals to the projectteam to train users, provide support, andadminister the software package. Many GISapplications, however, allow for in-housecustomization. Customized applications based onmacros or basic programming can significantlyreduce the learning curve required by engineersto use the software, as well as decreasedependence on a large staff of GIS analysts. Asingle GIS professional can create customizedmenu bars, tool buttons, and databases packagedwithin a clean and simple-to-use GIS add-onapplication. Well-thought-out training sessionscan further ease the transition to a new toolset.

The cost of creating a customized applicationshould not be a major stumbling block toimplementing a GIS-based infrastructuremanagement and design system. Currently,several commercial off-the-shelf GIS-baseddesign tools are available for wireless and outsideplant designs.

As demand for new outside plant networkscoincides with a shortage of trained and availableskilled personnel in this area, the GIS can beleveraged by network operators and engineeringfirms to fill that talent gap. Automatedspecialized GISs can allow users with lower skillsets and associated costs to assist the experiencedengineers already proficient in outside plantdesign. At the same time, the computer-savvyengineer often finds that using a GIS as an aidsignificantly reduces the learning curve foroutside plant design.

THE SPATIAL DATABASE

Beyond the many uses of a GIS intelecommunications applications, the greatest

power of a GIS lies in its ability to spatialize andintegrate databases. The basic data element of aGIS is a data table. Geographic features andattribute data alike are stored in flat tables similarto most existing database formats. It is widelyaccepted that 80 percent of all data has ageographic component. The ability to visuallyassess the locations of objects on the Earth’ssurface, rather then trying to interpret numberson spreadsheets, is a key element leading to theuse of a GIS in the first place.

Latitude and longitude coordinates can be storedin a central enterprise database side by side with infrastructure information, materials lists,previous designs, and imported client data. All of this content can then be layered togetherwithin the GIS to create a model of thetelecommunications infrastructure in question.The capabilities for SQL queries, compoundedwith geospatial functions, allow users to generate complex relational database solutions to geospatial questions. [9]

A shared corporate database releases engineersfrom the hunt for data (“Where is that list oftowers we built last year? Can we co-locate onthem?”). More often than not, the information isforgotten in someone’s desk drawer. Working,instead, in a database environment, engineers canaccess past projects and overlay that informationwith current project information. Networktopology, infrastructure, land base records, andother data are stored together on a singledatabase available to all users and all systems.Database structures such as the Spatial DataStandard for Facilities, Infrastructure, and theEnvironment (SDSFIE) are geared to the lowestcommon denominator. Drawings created in map-enabled CAD applications can be stored in thedatabase to be accessed by the GIS. Aerial sitephotos stored in the database can be overlaidwith existing utilities, past projects, and currentconstruction to create a complete picture of thecurrent situation. In this age of instant datagratification, the ability to gain immediate accessto all necessary information from one source to be used in one application simplifies theimplementation process.

CONCLUSIONS

While GISs have been used to great success in the wireless industry, their

full potential has not yet been reached in the telecommunications industry as a whole. The major GIS vendors are toutingtelecommunications applications and plug-ins forwireless and outside plant design andmaintenance. At the same time, majortelecommunications service providers withcustom-built legacy databases are being lockedinto dealing with specific contractors that arefamiliar with the software. The contractorsdealing with these legacy databases, which lackthe flexibility to integrate and analyze multipledata sources, are forced to jump through hoops towork with these systems and to reinvent thewheel project after project. In the process, they

The ability to visually assess

the locations of objects on theEarth’s surface,

rather then trying to interpret

numbers onspreadsheets,

is a key elementleading to the use

of a GIS.


stand the risk of stagnating on the integration ofGISs into their work and becoming increasinglyless competitive. Instead, they need to push notonly themselves to look to the future, but theirclients as well.

TRADEMARKS

Google Earth is a trademark of Google Inc.

Yahoo! is a registered trademark of Yahoo! Inc.

Microsoft is a registered trademark and VirtualEarth is a trademark of Microsoft Corporation inthe United States and other countries.

REFERENCES

[1] R. Paul, “Microsoft Launches Virtual Earth 3D to try to take on Google”(http://www.earthtimes.org/articles/show/10224.html).

[2] S. Smith, “AM/FM + GIS + Web,” GISVision magazine, December 1999.

[3] S. DuPlessis, “Geoinformation: A Singular Advantage in a Cellular Age”(http://www.geoplace.com/ge/2000/0500/0500gf.asp).

[4] K.C. Clarke, Analytical and Computer Cartography,Prentice Hall, 1995.

[5] B. Schweber, “With the Right Tools, You Can Score Big in the RF Field of Dreams,” EDN magazine, July 6, 2000.

[6] D. McCullough, “Four Options to Extend YourBroadband Service Revenues,” OSP magazine,October 2005.

[7] L. Godin, GIS in Telecommunications, ESRI Press, 2001.

[8] Personal interview with Bechtel networkengineer, October 2006.

[9] S. Rich, A. Das, and C. Kroot, “Spatial DataManagement in an Enterprise GIS”(http://gis.esri.com/library/userconf/proc01/professional/papers/pap742/p742.htm).

BIOGRAPHYPaul Lukas joined BechtelTelecommunications in 2004and during his first year created a customized fiber-opticnetwork design tool that fusedoutside plant design principleswith the geospatial datamanagement capabilities of aGIS. He has been the GISmanager for Bechtel Federal

Telecoms since 2005, where he evaluates, implements,and manages geospatial solutions. Paul also supportsthe Strategic Infrastructure Group and manages itsgeospatial data; he is currently developing a geospatialinfrastructure database and customized GIS applicationto expand the group's access to geospatial informationand solutions. Additional responsibilities have includedgeneral cartographic support, geospatial analysis forinfrastructure and telecommunications networks, andproposal support.

Before joining Bechtel, Paul worked for WirelessFacilities, Inc., where he provided dedicated geospatial support services for AT&T Wireless andassisted in creating a GIS-based wireless networkoptimization tool.

Paul began his studies in GIS at the Virginia PolytechnicUniversity and earned a BS in Technical Managementfrom DeVry University. He is a member of the Armed Forces Communications and ElectronicsAssociation (AFCEA).


INTRODUCTION

There are two basic means of providingcommunications services: wireless or wireline.

On the wireless side, the main hurdle isscarceness of radio frequency (RF) spectrum andthe associated huge cost. In the US, spectrum isviewed as a scarce national resource, closelyguarded by the Federal CommunicationsCommission (FCC). Based on the FCC’s personalcommunications services (PCS) auctions, themedian value of 1 MHz of spectrum per pop wasaround US$1.68 [1]. Simple math shows that a bare minimum of 10 MHz of spectrum (a pair of 5 MHz, enough for only one channel of current frequency division duplex [FDD]technologies such as universal mobiletelecommunications system [UMTS]) that covers 300 million US pops could cost close toUS$5 billion! And there is the cost of deployingthe network. On top of this, there are the ongoingsite rental or lease fees, which, on a nationwidebasis, could translate to hundreds of millions

or even billions of dollars annually. These factorsmake widespread usage of wireless broadbandrelatively difficult and expensive!

On the wireline side, there are currently twomeans of providing broadband services: digitalsubscriber line (DSL) through telephone companytelephone lines, and cable modem through cablecompany coaxial cable lines. Now, with theadvent of broadband over power lines (BPL orBoPL), a third wired option is emerging that useselectric utility power lines. Power lines areattractive for communications purposes becausethey have an omnipresence that reaches mosthomes and businesses, even in the most ruralareas. This ubiquity implies a possible reductionin both time and cost for broadband deployment.In this sense, power lines, like RF spectrum, canbe considered a very valuable national resource,or even a national treasure. And, of course, thereis the inside-home power line wiring that canliterally turn every outlet plug into a broadbandcommunications access port.

BROADBAND OVER POWER LINES (BPL)

Abstract—The Internet’s proliferation has focused attention on the importance of providing widespread access to broadband services. Many studies show that such access can have profound positive socioeconomicimpacts. Currently, however, broadband access is available to relatively few people worldwide. Broadband access has traditionally been provided via either DSL or cable. More recently, wireless and satellite broadbandaccess has also gained significant momentum. Now, a third—wired—option is emerging: broadband over power lines (BPL).

Power lines, however, were designed to deliver power, not communications, which poses three main hurdles forBPL. First, the variation in power line channel characteristics and performance over time and location must beappropriately considered. Second, measures must be put into place to ensure that BPL does not causeinterference for the existing rightful owners of the spectrum. Third, the regulatory issues accompanying anynew technology must be addressed.

As these hurdles are overcome, as standards mature, and as inexpensive standards-based equipment becomes more widely available, the concerns about the risks of BPL investment and deployment will gradually diminish.Then, the right business and deployment models will enable BPL to capture a significant portion of the thrivingbroadband market.

Key Words—access BPL, BPL, broadband over power lines, capacity, channel characteristics, coupler,extractor, FCC, injector, in-house BPL, interference, low voltage (LV) line, medium voltage (MV) line, noise,NTIA, Part 15, PLC, power line communications, repeater, Subpart G, transformer bypass


Lee [email protected]

S. Rasoul Safavian, PhD [email protected]



AC alternating current

AMR automated meter reading

AP access point

ARRL American Radio Relay League

AWGN additive white gaussian noise

BoPL broadband over power lines

BPL broadband over power lines

CALEA Communications Assistance for Law Enforcement Act

CENELEC European Committee for Electrotechnical Standardization

CFR Code of Federal Regulations (47 CFR addresses telecommunications)

CPE consumer premises equipment

CSMA/CA carrier sensing multiple access/collision avoidance

DAS distributed antenna system

dBm power in decibels with reference to 1 milliwatt

DSL digital subscriber line

EHV extremely high voltage (> 300 kV)

EM electromagnetic

EMC EM compatibility

EMI EM interference

ETSI European Telecommunications Standards Institute

FCC Federal Communications Commission

FDD frequency division duplex

FTTH fiber to the home

GDP gross domestic product

HDTV high definition television

HF high frequency (3 to 30 MHz)

HV high voltage (36 to 300 kV)

IEEE Institute of Electrical and Electronics Engineers

ISP Internet service provider

LAN local area network

LF low frequency

LV low voltage (< 1 kV)


MO&O Memorandum of Opinion & Order (FCC)

MTL multiconductor transmission line

MV medium voltage (1 to 36 kV)

NEC numerical EM code

NMS network management system

NOI Notice of Inquiry

NPRM Notice of Proposed Rule Making (FCC)

NTIA National Telecommunications and Information Administration

OFDM orthogonal frequency division multiplexing

OPERA Open PLC European Research Alliance

OSS operations support system

PC personal computer

PCS personal communications services

PL power line

PLC power line communications

POP point of presence

PSTN public switched telephone network

QoS quality of service

R&D research and development

R&O Report & Order (FCC)

RF radio frequency

RMS root mean square

ROI return on investment

SCADA supervisory control and data acquisition

SW shortwave (5.9 to 26.1 MHz)

UHF ultra high frequency

UMTS universal mobile telecommunications system

UPA Universal Powerline Association

USAC Universal Service Administrative Company

USF Universal Service Fund

UTC United Telecom Council

VHF very high frequency (30 to 300 MHz)

VoIP voice over Internet Protocol

Considering that broadband penetration iscurrently less than 4 percent globally, the hugegrowth potential for the broadband market isobvious. BPL could provide a quick and attractivesolution. Of course, successful BPL deploymentrequires not only a solid technical performanceand field trial records, but also realistic and viablebusiness and deployment plans.

This paper first examines the current state ofbroadband access and the importance of havingthis access. Then, a quick overview of the electricpower grid and how it can be altered to allowBPL sets the stage for a review of the current BPLplayers, field trials, commercial deployments,and standards bodies. This is followed by a briefexamination of the potential benefits of BPL to the electric utility companies, service providers,and end-users and a look at the main challenges for BPL, namely harsh power linechannel characteristics and performance issues,interference concerns, and the regulatoryactivities surrounding BPL. The paper continueswith a review of the BPL business models andeconomic issues before presenting conclusionsand closing remarks.

BROADBAND ACCESS

Current State of AccessDespite the widespread and spectacular growthof broadband technologies in the last few years,significant regions of the world, including ruraland low income areas in the US, still do not haveaccess to broadband services. In fact, out of the 6.7 billion people who currently inhabit ourplanet, roughly 3.7 billion (60 percent) haveaccess to electrical power services, whereas onlyabout 2 billion (30 percent) have access to sometype of telephony services (wireline and/orwireless), and only roughly 250 million (3.7 percent) have access to broadband services [2, 3]. In the US, out of a population of 300 million—and using a relaxed definition of broadband as only 200 kbps in at least onedirection (Internet to user [receiving or downlink]or user to Internet [transmitting or uplink])—onlyroughly 50 million people currently have accessto broadband services.

A major hurdle to deploying broadband servicesis the high cost of deploying the so-called last-mileaccess. The last mile (also sometimes referred to as the first mile, local loop, or access network)is defined as the part of the network that links users with broadband services. From acommunications perspective, power lines, due to

their omnipresence and the fact that they havealready reached electrical power users in homesand offices, would seem to solve this access issue.In this sense, they may be considered as apossible third set of broadband wires reachinghomes or businesses (the other two being DSL and cable modem). Of course, last-milebroadband access could also be providedwirelessly via fixed wireless, cellular, or satellite systems.

The wiring inside a home or office can also beused to provide a local area network (LAN)connecting computers, printers, and smartappliances and basically turning every outlet intoan Internet connection. This is sometimes referredto as last-inch access or connectivity.

It is worth noting that while industrializedcountries typically have several—albeitsometimes prohibitively pricey—telephony andbroadband options, less developed countries mayhave access only to power line services andfrequently lack well-established conventionaltelecommunications infrastructure. It is here thatpower line communications can be particularlyuseful and effective. Households connected topower lines may be quickly provided withtelephony via voice over Internet Protocol (VoIP)and broadband Internet services, with minimalneed for a new major infrastructure and itsassociated huge financial investment. For manyof those underserved communities, this would betheir first access to telephony, Internet, andrelated services.

Importance of AccessNumerous studies have shown a directrelationship between the availability andpenetration rate of broadband and animprovement in productivity, quality ofeducation, quality of health care, generation ofnew high-paying jobs, and facilitation of newchannels for commerce. These, in turn, can alllead directly to national economic growth (with adirect impact on gross domestic product [GDP])and even enhanced national security. Accordingto Thomas L. Friedman, the frequently quotedop-ed commentator on globalization:

Jobs, knowledge use and economic growthwill gravitate to those societies that are themost connected, with the most networks andthe broadest amount of bandwidth—becausethese countries find it easiest to amass,deploy and share knowledge in order todesign, invent, manufacture, sell, provideservices, communicate, educate and enter-tain. Connectivity is now productivity. [4]


Considering thatbroadband

penetration iscurrently less than 4 percent globally,

the huge growthpotential for the

broadband marketis obvious.

BPL could provide a quick and

attractive solution.


Unfortunately, both nationally and globally, alarge digital divide, or gap, separates those with regular and effective access to digitaltechnologies and those without. Morespecifically, a gap exists between people whohave effective high-speed Internet access (theinformation haves) and those who do not (theinformation have-nots). Realizing the importanceof broadband, US President George W. Bush, onApril 26, 2004, called for providing universal andaffordable broadband access in every part ofAmerica by 2007 as part of his initiative to create“A New Generation of American Innovation” [5].

With respect to the president’s broadbandinitiative, BPL could play an important role by offering:

• Affordability: With no need for new wiringor major infrastructure deployment, BPLcreates an alternative broadband solutionthat could lead to lower prices for broadbandconsumers.

• Universality: BPL could facilitate and speedup connecting the rural and low income parts of America to broadband services,thereby helping to bridge the digital divide.

Thus, power lines could perform double duty bydelivering electrical power services andproviding broadband information services. BPLdeployment, in turn, holds the promise of

providing both telephony (via VoIP) andbroadband services to all 3.7 billion people on our planet who have access to power lines!

It is also worth mentioning that power linecommunications (PLC) is not a new subject, butone that has been around for decades. Severalpower companies around the globe have beenusing power lines for low-speed applications (a few kbps in the low frequency [LF] portion ofthe spectrum), such as power line metering and control. The recent renewed interest in using power lines for communications revolvesspecifically around providing BPL applications.The main idea is to use specialized equipment to slightly modify the existing power grid toallow it to also carry high speed data over a broad spectrum range (high frequency [HF], the lower portion of very high frequency [VHF], and potentially beyond) without causingunreasonable interference to the rightfulincumbent users of those RF bands. Furthermore,this has to be done in an economically andfinancially viable manner.

ELECTRIC POWER GRID

Overview of Grid Structure and TopologyWhile the details of electric power grid structuresand topologies differ from country to country,

High Voltage Transmission Lines

Medium Voltage Transmission Lines

Power Substation

Power Plant

Power Substation

Low Voltage Transmission Lines

Low Voltage Transmission Lines

Medium Voltage Transmission Lines

High Voltage Transmission (69 kV and Above)

Primary Distribution Medium Voltage (2.4 to 35 kV)

Secondary Distribution Low Voltage (Up to 600 V)

Figure 1. Typical Electric Power Grid

US PresidentGeorge W. Bush,

in April 2004, called for providing

universal andaffordable

broadband accessin every part of

America by 2007 as part of his

initiative to create“A New Generation

of AmericanInnovation.”


a power grid basically consists of power plants or generators, transmission substations,transmission lines, power substations withtransformers to change voltage levels, anddistribution lines that collectively generate and carry the electricity from power plants all the way to wall plugs. See Figure 1.

Power plants are basically spinning electricitygenerators. Spinning can be performed by asteam turbine, and steam can be created byburning fossil fuel or from a nuclear reactor. A generator’s output is three-phase alternatingcurrent (AC) power at voltage levels in thethousands. The three single phases aresynchronized and offset by 120 degrees. Three-phase current is chosen because single-phase AC goes through a full cycle (from zero to peak to zero to other peak and back to zero) at the line rate, which is 60 times per second in the US and 50 in the other parts of the world. With three synchronized phases, onthe other hand, one of the three phases is nearinga peak at any given instant. More phases could beused, but this implies more wires and higher cost;three seems to be a good compromise betweencost and performance.

Power P, transferred over lines and delivered tocustomers, is equal to the product of voltage Vand current I (P = IV). Power loss in the linegrows with the square of the current, that is, Ploss = Rline • I2, where Rline is the line resistanceand depends on the line material and increaseswith the length of the line. For a given generatedP and a given Rline , to reduce Ploss , current I mustbe made as small as possible. This means that the line voltage must be made as large as possible,especially for long-distance transmissions.Transmission substations located next to powerplants use large transformers to step up generatoroutput from thousands of volts to hundreds of

thousands of volts (typically between 155,000 and765,000 volts), thus allowing megawatts of powertransmission over distances of 300 miles or more.

At power substations, voltages are stepped downand lines are branched out to cover larger areas.This is performed successively, transforming andbranching out from extremely high voltage (EHV,typically 155 to 765 kV) to high voltage (HV,typically 45 to 155 kV), and then from HV tomedium voltage (MV, typically 2 to 45 kV), andfinally from MV to low voltage (LV, typically 100 to 600 V) for delivery to homes or businesses.The result is a tree-structured power distributionhierarchy. Basically, EHV and HV are used totransmit AC electric power, and MV and LV areused to distribute it. See Figure 2.

The structures needed to support EHV and HVlines are typically tall, massive towers. MV andLV lines, on the other hand, are typicallymounted on street poles. In the US, street polesare typically 10 meters high, located 50 metersapart, and support three wires that carry the three separate phases, plus a neutral (possiblygrounded) wire. A network of MV lines is usuallyreferred to as the primary distribution; a networkof LV lines is the secondary distribution.

In the US, at the primary distribution level, mostpower lines are aerial or overhead. At thesecondary distribution level, particularly innewer urban areas, most lines run underground.Overhead lines are more susceptible thanunderground lines to producing radiationinterference and to picking up interference. Butunderground lines are used less due to theprohibitive cost of burying cables. In the US, MV lines typically run between 15 and 50 km.

As mentioned, levels and structures of branching,network architectures, and voltage levels varyfrom country to country. For instance, in the US,

TransformerGeneration

MV

EHV HV MV

MV

MV

MVHV

HV

Transformer

Transmission Distribution

LV

LV LVLV

LV

LV LVLV

LVConsumption

Figure 2. From Generation to Consumption: Power Grid Hierarchies

A power gridbasically consists

of power plants or generators,transmissionsubstations,

transmission lines,power substationswith transformers

to change voltage levels, anddistribution linesthat collectively

generate and carry the electricityfrom power plants

all the way to wall plugs.


typically fewer than a dozen homes are served by a single MV/LV transformer, whereas inJapan this number is about 30 and in Europe it is several hundred. This affects not only thecommunications characteristics, but also theeconomic viability of a BPL system. (BPL businessmodels are examined later in this paper.)

Altering the Power Grid To Allow BPLEHV and HV lines are usually too noisy totransmit broadband communications signals;only MV and LV lines are used for BPL. MV lines are usually less branched than LV lines,making point-to-point connections possible. MV networks allow communication over longer distances because of their weaker signalattenuation and lower noise level.

To use power lines for broadband communi-cations, the broadband signal must be injectedinto and extracted from the lines throughcouplers. LV couplers may be capacitive orinductive, depending on distribution systemtopology, performance requirements, and cost. Incapacitive coupling, a capacitor is responsible for

the actual coupling, and the signal is modulatedonto the network’s voltage waveform. Ininductive coupling, an inductor is used to couplethe signal onto the network’s current waveform.Inductive couplers are known to be rather lossy,but since they require no physical connection tothe network, they are safer to install on energizedlines than capacitive couplers. MV couplers aretypically inductive. It is important that couplersbe easy-to-install passive devices with low failurerates that can be used outdoors and installed onenergized lines.

Line noise, limitations on the amount of signalpower that can be injected into power lineswithout causing unacceptable interference forother spectrum users, and signal attenuation asthe signal traverses the line make it necessary toregenerate or repeat the signal periodically. Thiscan be done by using MV couplers to couple thebroadband signal off of the MV line so that it canbe regenerated if necessary and amplified beforebeing fed back onto the MV line through anothercoupler. Repeaters, on the other hand, could addlatency (especially if the signal is regenerated)

Mobile Network PSTN

BackhaulBox

Power Substation

Backhaul Network

Internet

Access BPL In-House BPL

Power Generator

MV Lines

MV Coupler

MV Lines

HV Transmission Lines

Repeater Box

Transfer Bypass

Box

MV Coupler

LV CouplerMV

Coupler

LV Lines

PCVoIP

Phone

MV Lines

LV Lines to Homes/Businesses

MV Line

MV Coupler

LV Coupler

LV Line to Home/Business

Transfer Bypass

Box

Transformer

LV Line

Coax

Figure 3. Typical BPL Architecture

Couplers should beeasy-to-install

passive devices with low failure

rates that can be used

outdoors andinstalled on

energized lines.


and could also create single points of failure,because a single bad repeater can bring down anentire communications line.

The distribution transformers that change voltage levels between MV and LV lines areparticularly harsh on the weak broadband signal.Transformers, which are intended to pass lowfrequencies near 50 or 60 Hz, appear as opencircuits for the passage of higher frequencysignals and typically attenuate and distort theweak broadband signal beyond reconstructionand usability. This implies that BPL signals going between MV and LV lines need to bypassthe transformers. Typically, the bypass box canalso have built-in repeating functionality at asmall incremental cost. The recent capability toeffectively and safely bypass transformers hasbeen instrumental to the success and deploymentof BPL.

A point-of-presence (POP) is needed to connectthe BPL network to a backhaul network such as the Internet, a public switched telephonenetwork (PSTN), or a mobile network. Theconnection is made through a backhaul networkbox coupled to an MV distribution line, typicallynext to a power substation where multiple MV lines are connected. The backhaul networkbox is typically a bidirectional device thatconverts data formats, aggregates andconcentrates uplink data streams, providesrouting functionality, helps allocate bandwidthand resources, generates billing and chargingdata, and provides various backhaul Ethernetinterfaces to fiber optic or wireless connections.Figure 3 illustrates a typical BPL architecture.

A BPL network, like any other communicationsnetwork, also requires a network management

system (NMS) or operations support system(OSS) to observe and manage network resourcesand perform billing and other back-end tasks.

BPL Deployment Options The MV and LV line portions of the BPL areusually referred to as the access BPL, while theportion inside a home or office using the insidewiring is called the in-house BPL. BPL can bedeployed either as end-to-end BPL or as hybridBPL, using one of the three options illustrated in Figure 4.

An end-to-end BPL system uses both access BPL and in-house BPL, i.e., power lines are usedall the way from the power substation to the end user. Two of the three BPL deploymentoptions involve the access BPL portion of an end-to-end system: the BPL signal can either (1) bypass the MV/LV transformer (as doesCURRENT Technologies® equipment) or (2) gothrough the transformer (as does MainNetCommunications equipment).

The third BPL deployment option is hybrid BPL.In this option, typically only the MV lines areused, and a fixed wireless network replaces theLV lines and in-house BPL (Amperion™ takesthis approach). In hybrid BPL, the bypass boxdoes not couple the broadband signal to/from the LV line but converts it to/from a wirelessformat and delivers it to the wireless access point(AP) also located on the pole.

These different deployment options have theirassociated performance and cost tradeoffs. Forend-to-end BPL, bypass boxes and LV couplersmust be installed on all LV lines, and in-houseBPL modems are required. For hybrid BPL,bypass boxes with wireless conversion boards,

The MV and LV lineportions of the BPL

are usually referred to as

the access BPL,while the portioninside a home oroffice using the

inside wiring is called the

in-house BPL. BPL can be

deployed either as end-to-end BPLor as hybrid BPL .

Substationwith Modem

Injector

Option 1Transformer

Bypass System

Option 3Wireless Connection

Option 2Through

Transformer

Repeater

Extractor

Coupler

Coupler

Router

Wireless Transmitter

with AntennaWireless Receiver

with Antenna

Figure 4. BPL Deployment Options


wireless APs, and existing standard wireless user modems are required, but LV transformerbypasses and LV couplers are not. Alsoassociated with hybrid BPL are the usual existing issues regarding wireless performance in unlicensed spectrum and the current state of wireless quality of service (QoS), security, and so forth.

INDUSTRY PLAYERS, FIELD TRIALS,COMMERCIAL DEPLOYMENTS, AND STANDARDS BODIES

Industry Players, Field Trials, and Commercial DeploymentsGlobally, the number of BPL players (electricutility companies, equipment manufacturers,investors, etc.), field trials, and commercialdeployments has been growing steadily in thelast few years. In the US alone, there have beenmore than 39 trial deployments [6]. CURRENTTechnologies is currently offering commercialBPL services with Duke Energy in Cincinnati,Ohio, with plans to expand elsewhere withinDuke’s 1.5-million-customer service territory inOhio, Indiana, and Kentucky. CURRENTTechnologies is also planning to deploy BPLservices to potentially 2 million residents ofDallas, Texas, using TXU Electric delivery. TheCity of Manassas, Virginia, has been offeringcitywide BPL services using MainNet equipmentsince 2005. Progress Energy and EarthLink® planto provide BPL services in North Carolina usingAmperion equipment.

There are also commercial deployments in Spain,Germany, Korea, Chile, Brazil, and the UK. In Spain, Endesa began service in 2003 in Saragossa and Barcelona; Iberdrola initiatedservice in Madrid and Valencia in the same year.Power Plus Communications has started offeringservices in Germany, as has Scottish SouthernElectric in the UK.

Standards BodiesStandardization is of paramount importance to the success of any new technology such as BPL. To this end, the Open PLC European Research Alliance (OPERA), EuropeanTelecommunications Standards Institute (ETSI),Institute of Electrical and Electronics Engineers(IEEE), Universal Powerline Association (UPA),European Committee for ElectrotechnicalStandardization (CENELEC), and HomePlug®

Powerline Alliance have been leading theactivities and creating their own standards.

OPERA—a consortium of currently 37organizations, including electric utilitycompanies, PLC equipment manufacturers, and universities—is a research and development(R&D) project with funding from the EuropeanCommission to create and promote open globalspecifications for low-cost, high-performance,high-speed power line communications. Its first specification documents were released onFebruary 21, 2006. These specifications will bepromoted through international standardizationorganizations, including IEEE and ETSI [7].

The IEEE BPL study group drove the creation of the BPL-related Pxxxx working groups. The IEEE P1675™ “Standard for Broadband over Power Line Hardware” Working Group is chartered to develop standards for power line hardware installation and safety. The IEEE P1775 “Powerline CommunicationEquipment – Electromagnetic Compatibility(EMC) Requirements – Testing and MeasurementMethods” Working Group is focused on PLC equipment, electromagnetic compatibilityrequirements, and testing and measurementmethods. The IEEE P1901 “Draft Standard for Broadband over Power Line Networks:Medium Access Control and Physical LayerSpecifications” Working Group is responsible fordefining the medium access control (MAC) andphysical layers for high speed (greater than 100 Mbps at the physical layer) for both in-house and access BPL. The standard will focuson transmission frequencies below 100 MHz. The specifications of these working groups arescheduled for release in 2007 [8].

The UPA has also released a number ofspecifications related to different aspects ofpower line technology. Three main specificationsare the UPA coexistence specification, released inJune 2005; the UPA access BPL specification,endorsed by OPERA and released in February 2006; and the UPA in-house BPLspecification, called Digital Home Standard v1.0and also released in February 2006. The UPA also works with and through internationalstandardization bodies such as IEEE and ETSI topromote its standards [9].

The HomePlug Powerline Alliance was foundedin 2000 and currently has over 65 membercompanies. The alliance’s standards (HomePlug1.0 and AV) are for home networking over power lines (in-house BPL). The HomePlug 1.0specification allows for speeds up to 14 Mbps.The current HomePlug AV specification allowsfor speeds greater than 100 Mbps (suitable for

Globally, the number of

BPL players (electricutility companies,

equipmentmanufacturers,investors, etc.), field trials, and

commercialdeployments

has been growingsteadily in

the last few years.


high definition television [HDTV] and VoIP) andis compatible with HomePlug 1.0. In 2004, toprovide a harmonized end-to-end BPL standard,the HomePlug Powerline Alliance started lookinginto creating an access BPL standard planned forcompletion by early 2007 [10].

POTENTIAL BENEFITS

Benefits to Service ProvidersFrom a service provider’s point of view, BPLcould provide large cost savings. The first, and byfar the most important, factor is that thetransmission medium, i.e., the power lines, isalready in place. There is no need to purchasespectrum or to hang, dig, or lay new wires,because most of the required infrastructurealready exists. There is also no need for thedifficult, expensive, and time-consuming siteacquisition, permitting, and licensing tasksneeded for a typical deployment. Given theomnipresence of power lines, BPL also holds thepromise of being able to provide genuinelyubiquitous coverage. These factors implypotential cost and time savings that could levelthe BPL deployment playing field a bit morecompared with DSL and cable, both of whichhave significant deployment head starts.

Benefits to Electric UtilitiesFor the electric utility companies, BPL’s benefitsare twofold: (1) It can create new sources ofrevenue from an existing investment, and (2) it can help create a smart grid for the utilitycompanies that would enable enhanced utilityapplications [11, 12] such as:

• System monitoring from any point on theelectric grid

• Load shifting and balancing

• Optimized asset utilization and management

• Performance of preventive maintenance andimprovement of service reliability andcustomer satisfaction by avoiding poweroutages and emergencies

• Advanced supervisory control and dataacquisition (SCADA)

• Fault detection, fault analysis, and adaptiveself-healing

• Automatic outage detection, restorationdetection, and verification

• BPL-enabled electricity meters that enabletime-of-day and real-time pricing through

automated meter reading (AMR) with remote disconnect (and reconnect) and theft detection

• Real-time video surveillance of the sensitivenational power infrastructure (e.g., grid and substations)

Benefits to End UsersEnd users can benefit from BPL deploymentbecause:

• BPL could create competition and thus helpreduce end-user service prices.

• BPL could provide high user throughputs, as discussed later in this paper.

• In some places, BPL may be the only viablechoice (e.g., in rural areas), although satellite-based service may also be of interest in these areas.

• BPL could be used for smart appliances,connected and controlled through a PC andremotely. While these devices could possiblybe controlled through a DSL or a cablemodem connection, BPL may provide a moreintegrated (neater) solution.

• BPL may provide a more ubiquitous andreliable service coverage area.

The explosive growth of the Internet and therecent deregulation of telecommunications in theUS and Europe have led to the renewed interestin BPL. Extensive research on BPL channelmodeling [13–20] and a considerable amount ofinterference analysis [21–25] have taken place.Concurrently, there have been a large number of field trials and measurements to validatevarious models [21–31], along with advances insignal processing such as the newer adaptivemodulation and coding techniques [28] andfaster, cheaper processors and electronics.Nonetheless, despite its renewed attractiveness,BPL must overcome implementation challengesas well as regulatory concerns before it canbecome a viable avenue of broadband access. Thenext sections of this paper examine in more detail the key implementation challenges andregulatory concerns facing BPL.

IMPLEMENTATION CHALLENGES

The Nature of the Power GridThe most obvious challenges to implementingBPL arise from the fact that power line grids wereoriginally developed to transmit electrical power

The explosivegrowth of theInternet and

the recentderegulation of

telecommunicationsin the US and

Europe have led tothe renewed

interest in BPL .


(high voltage AC at low frequencies of 50 or 60 Hz)from a small number of sources (the generators)to a large number of sinks (the end customers).Power grids were neither designed nor devisedfor communications purposes. Even though theinterest in using power lines for communicationsis not new, their early use for data transmissionwas mainly for simple, low-data-rate (a fewkilobits per second) remote monitoring and meterreading applications at a low frequency (typicallyonly up to a few hundred kilohertz).

The main challenges to BPL arising from thenature of the power grid have been the extremelyharsh, unpredictable, time-and-location-variablecharacteristics of the power line channel, and potential interference concerns (in bothdirections) [13–25]. Because power lines are nottwisted and have no shielding, they can produceelectromagnetic radiation that is easily detectedby radio receivers. For the same reasons, powerlines can also easily pick up nearby radiofrequency signals. Thus, addressing mutualinterference is not only a challenge, but becomesa valid regulatory concern.

A related challenge facing BPL centers arounddata sensitivity. To prevent interception ofsensitive data by unintended and unauthorizedreceivers, data encryption is a must.

The fact that the power line grid is a sharedmedium and BPL is a contention-based systemcreates additional challenges. Because all usersshare the available channel capacity orbandwidth, as the number of users goes up, per-user throughput goes down. In the US, thereare typically 50 homes per substation. An averageavailable throughput of 50 Mbps implies roughlyan average of 1 Mbps per user, a speed on parwith the current average speeds delivered by DSLor cable modem. However, BPL is thought to bedistance limited, similar to DSL. Thus, thedistance between the customer’s home and thesupplying substation is a factor in the bit rate available to the user.

Channel Characteristics and Capacity

Power Line NoiseIn general, a power line channel is a very harsh and noisy transmission medium. The noise on the line is typically time, location, andfrequency dependent.

Time-variable behavior is due mainly to thedynamically changing nature of the loadconnected to the power lines. Line branching, thenumber and types of branches, the lengths of line

segments, the types of power line equipmentconnected (such as capacitor banks andtransformers), and the kinds of loads connectedall affect channel characteristics. Furthermore,impedance mismatches caused by unterminatedstubs and line branches cause signal reflectionsand create a frequency-dependent fadingchannel, much like the multipaths typically seenin mobile wireless communication channels.

MV and LV lines have very different noisecharacteristics. The MV grid is usually lessbranched than the LV grid, and LV lines aretypically terminated at time-varying consumerelectrical appliances. Noise on the LV grid istypically the sum of background noise, impulsivenoise, and synchronous/nonsynchronous (withthe power line frequency) colored noise,generated primarily by electrical appliances; thisnoise is certainly not an additive white gaussiannoise (AWGN). On the MV grid, the on/offswitching of the capacitor banks used to correctthe power factor typically causes high noisepeaks [14]. At the same time, background noiseand narrow-band noise are dominant on MVlines. The background noise is environmentalnoise that is highly dependent on weather,location, and elevation. The narrow-band noise is caused by RF interferers such as amateur orshortwave (SW) radios and varies randomlyacross location and time. Noise levels on MV lines are typically as much as 20 to 30 dBhigher than on LV lines in the frequency range of 1 to 20 MHz [21].

Channel AttenuationPower lines have been modeled in the literatureby using either statistical approaches based on extensive measurements or deterministicapproaches based on multiconductor trans-mission line (MTL) theory and numericalanalysis. Carson’s earlier MTL model [17]allowed for ground impedance but did notinclude ground admittance, which cannot beignored in higher frequencies and/or under poor conductive ground plane conditions. The subsequent MTL models in [18, 19] includeground admittance.

A simple matched uniform MV line segment withno connected device or junctions could have aslittle as 1 dB/km ohmic absorption or attenuationloss. For a complex overhead MV network, on theother hand, the amplitude of the channelfrequency response (or, equivalently, the channelattenuation) in the frequency range of 10 kHz to100 MHz shows highly frequency-dependentattenuations of as high as 40 dB/km caused byreflections from abrupt discontinuities and

The main challengesto BPL arising from

the nature of the power grid

have been the extremely harsh,

unpredictable, time-and-

location-variablecharacteristics of

the power linechannel,

and potentialinterference

concerns (in bothdirections).


mismatched impedances [23]. LV network lossesare typically higher than MV network losses andcould be as high as 100 dB/km [14].

Performance ImprovementsConditioning the grid can improve power line performance by minimizing impedancemismatches, terminating stubs, filtering noise,etc. These options, however, may deteriorate ordiminish the advantages of power line grids. Abetter approach is to use modulation and codingschemes robust enough to work in the hostilepower line channel environment. Currently, mostBPL products use orthogonal frequency divisionmultiplexing (OFDM), well known for itsexcellent robustness against channel distortionssuch as multipath and impulsive noise and for itsgood spectral efficiency, reasonable cost, andability to avoid certain bands.

In BPL systems, multiple user modems areconnected in a bus or star topology. Some type of MAC must be implemented to providecommunications through shared bandwidth onpower lines. To provide the necessary QoS for applications that require bandwidth and performance guarantees, such as videostreaming, the carrier sensing multipleaccess/collision avoidance (CSMA/CA) protocolmay be used. This widely used scalable protocol, also used in the wireless fidelity (IEEE 802.11) MAC layer, is suitable for powerline channel characteristics.

Capacity and Spectral EfficiencyDepending on the bandwidth used on the powerlines (typically a frequency range between 2 and100 MHz), on the BPL injection power level(typically 1 to 30 dBm), and on load and channelconditions, throughputs in the range of tens, oreven hundreds, of megabits per second andspectral efficiencies in the range of 1 to 20 bps/Hzcan be achieved [20]. Theoretical and field trialshave also claimed throughputs of the same orderof magnitude, and even in the gigabit-per-secondrange if larger frequency bandwidths in the upper VHF/ultra high frequency (UHF)spectrum and higher input signal powers areused. In the US, however, this may not be a viableoption, considering that licensed spectrum in theVHF and UHF bands is heavily occupied.

A system developed by Corridor Systems, Inc., inthe US uses MV power lines in frequency rangesfrom VHF through microwave as distributedantenna systems (DASs) to extend existingcellular network coverage [29]. The cellularnetwork RF signal is picked up by the Corridor

equipment, converted into a proprietary BPLformat, and injected into and transported downthe MV lines. At cellular dead zones, the Corridorequipment converts the signal back to its originalformat for re-radiation by local antennas. Thus,MV lines are used to carry cellular signals to areastoo difficult or expensive to reach by cellularnetworks, conventional repeaters, or DASs.

Interference Concerns and Regulatory IssuesUnlike the twisted wires of telephone companiesand the shielded cables of cable companies, longunshielded, untwisted, overhead power lines canact as large antennas and be natural sources andtargets of electromagnetic interference (EMI). Inaddition, BPL signals tend to radiate from theinjectors and repeaters spaced along the powerlines. This raises concerns about interfering with the rightful owners of the radio spectrum inthe BPL range of operation [30]. The mostconcerned and vocal opponents of BPL in the US are amateur radio operators, through theAmerican Radio Relay League (ARRL), andgovernment agencies.

The US FCC started examining the use of powerlines for broadband communications services byissuing a Notice of Inquiry (NOI) on April 23,2003. The NOI sought information on potentialinterference from BPL systems and associatedchanges that may be needed to accommodate BPLsystems in Part 15 of the FCC’s rules published in the Telecommunications Code of FederalRegulations (47 CFR).

Part 15 addresses RF devices. Part 15, Subpart A,addresses general issues. Section 15.3 definesterms used in the FCC’s rules. Subpart Baddresses unintentional radiators, with Section15.109 defining the radiated emission limits.Subpart C deals with intentional radiators, withSection 15.209 defining the corresponding generalrequirements and radiated emission limits.Section 15.3 (f) defines a carrier current system asa system, or part of a system, that transmits RFenergy by conduction over electric power lines.

A carrier current system can be designed so thatthe RF signals are received by conduction directlyfrom the connection to the electric power lines(unintentional radiator) or so that the signals arereceived as over-the-air radiation from theelectric power lines (intentional radiator). Carriercurrent systems operate on an unlicensed basisunder Part 15. As a general condition ofoperation, Part 15 devices may not cause harmfulinterference to authorized radio services andmust accept any interference they receive.

Unlike the twisted wires of

telephonecompanies and

the shielded cablesof cable companies,

long unshielded,untwisted, overheadpower lines can actas large antennas

and be natural sources

and targets of EMI.


The FCC amended the existing Section 15.3 toinclude Sections 15.3 (ff) for access BPL and 15.3 (gg) for in-house BPL, as follows:

Section 15.3 (ff) – Access BPL: A carriercurrent system installed and operated on anelectric utility service as an unintentionalradiator that sends radio frequency energy on frequencies between 1.705 MHz and 80 MHz over medium voltage lines or overlow voltage lines to provide broadbandcommunications and is located on the supply side of the utility service’s points ofinterconnection with customer premises.

Section 15.3 (gg) – In-House BPL: A carriercurrent system, operating as an unintentionalradiator, that sends radio frequency energyby conduction over electric power lines thatare not owned, operated or controlled by anelectric service provider. The electric powerlines may be aerial (overhead), underground,or inside the walls, floors or ceilings of user premises. In-House BPL devices mayestablish closed networks within a user’spremises or provide connections to AccessBPL networks, or both.

In its response to the FCC’s NOI, the National Telecommunications and InformationAdministration (NTIA) of the US Department of Commerce described the federal government’susage of the 1.7 to 80 MHz spectrum, identifiedassociated interference concerns, and outlined thestudies it planned to conduct to address thoseconcerns. In April 2004, the NTIA published its Phase 1 Study technical report, NTIA Report 04-413, “Potential Interference fromBroadband over Power Line (BPL) Systems toFederal Government Radiocommunications at 1.7 – 80 MHz” [31]. In this report, the NTIAdefined interference risks to radio reception in theimmediate vicinity of overhead power lines used by an access BPL system. The radio systemsto be considered in interference analyses includeda land vehicular receiver, a ship-borne receiver, a receiver using a rooftop antenna (e.g., a base or fixed-service station), and an aircraft receiverin flight. The study included variousmeasurement campaigns and the use ofnumerical electromagnetic code (NEC) softwareto characterize BPL signal radiation andpropagation and to evaluate interference risks.The report also suggested means for reducinginterference risks and identified techniques formitigating local interference should it occur. The Phase 1 Study focused on simple BPLdeployment models. The Phase 2 Study isfocusing on evaluating the effectiveness of theNTIA’s Phase 1 recommendations and on theresults of a study of potential interference via

ionospheric propagation of BPL emissionsresulting from the mature large-scale deploymentof BPL networks. As of the date of this paper, thePhase 2 Study report had not yet been released.

Some of the NTIA’s Phase 1 Study highlightsinclude:

• In the 1.7 to 80 MHz spectrum, the dominantpropagation modes are ground waves, spacewaves, and sky waves. Ground waves consistof direct waves, ground-reflected waves, andsurface waves. Direct waves decay at a rateproportional to the square of their distancefrom their source. Ground-reflected waves(along with direct waves) decay at the rate ofdistance raised to the power of four. Ground-reflected waves may be of no major concern ifthe radiator is relatively far from ground.Surface waves propagate close to the groundand have a substantially higher rate ofattenuation than direct waves. Ground wavepropagation is pertinent on BPL signal pathsbelow the power line horizon. Space wavesinvolve only direct waves and occur overelevated signal paths, e.g., signal paths abovethe power line horizon. Sky waves areparticularly important in the HF band (forBPL, 1.7 to 30 MHz) and have temporal andspatial variability. Here, signal paths arerepresented as rays reflected and refracted bythe ionosphere. Sky waves can extend thesignal’s reach to several kilometers.

• The space around a radiator is typicallydivided into three regions: reactive near-field, radiating near-field, and far-field.These regions are typically defined as:

where r is the distance from the radiator, D isthe largest linear dimension of the radiator,and λ is wavelength. For BPL systems, thevictim receiver is typically in the radiatingnear-fields, although far-fields are importantbecause of sky waves and at distances seenby aircraft receivers.

λD3

r < 0.62

λD3

0.62 < r < 2 λD2

r > 2 λD2

Reactive Near-Field

Radiating Near-Field

Far-Field

The FCC amendedthe existing

Section 15.3 to include

Sections 15.3 (ff)for access BPL and 15.3 (gg)

for in-house BPL .


The NTIA also provided some recommendationsand suggested some interference mitigationtechniques; these include:

• Mandatory registration of certain parametersof planned and deployed BPL systems

• A requirement for BPL devices to befrequency agile (i.e., to have notching andretuning capabilities) and to have remotepower reduction and shutdown capabilitiesto eliminate interference if any is reported

• Use of minimal required power

• Avoidance of locally used radio frequencies

• Use of symmetry and differential modesignal injection to minimize radiation [31, 23].Symmetry is defined in terms of impedancebetween conductors and ground. If, for atwo-wire line, the impedance between eachconductor and ground is equal, the line issymmetrical or balanced. Balanced lines arenecessary for differential mode transmission,in which the currents are equal in magnitudeand flow in opposite directions on theconductors. The fields radiating from theseconductors tend to cancel each other.

Subsequent to the above activities, the FCCreleased its Notice of Public Rule Making(NPRM) in February 2004, and received morethan a thousand comments and replies frommany concerned parties [32]. The FCC eventuallyfinalized its decision by adopting its Report &Order (R&O) FCC 04-245 on October 14, 2004(published in the Federal Register on January 7,2005) [33]. The FCC considered various petitionsto reconsider the R&O and subsequentlyamended the Part 15 rules to modify some of theprevious specified exclusion zones and add a fewnew exclusion zones. However, the FCC deniedother petitions to reconsider other aspects andpublished the final Memorandum Opinion &Order (MO&O) on August 7, 2006, and the newamended rules in 47 CFR.

The FCC basically decided to keep BPL underexisting Part 15 unlicensed device rules andadded Subpart G for access BPL. Morespecifically, Sections 15.601, 15.607, 15.611, and15.613 of this new Subpart include the followingnew rules:

• Exclusion Bands: These are certain bands of frequencies within which access BPLoperations are not permitted.

• Exclusion Zones: These are certaingeographic areas within which access BPL operations are not allowed.

• Consultation: A consultation is to be heldbetween an entity operating access BPL and alicensed public safety or other designatedpoint of contact, for the purpose of avoidingpotential harmful interference.

• Equipment Authorization: Because BPL is anew technology, the FCC has required thatall BPL-related equipment be certified.Certification is an equipment authorization bythe FCC or its designated entities, as opposedto verification, which is a manufacturer’s self-approval procedure. The rules adopted in theR&O require that all access BPL devicesmanufactured, imported, marketed, orinstalled 18 months or later after the FederalRegister publication of the R&O (i.e., afterJuly 7, 2006) must comply with the newly adopted requirements of Subpart G of Part 15 for BPL devices, includingcertification of the equipment.

• Databases: Publicly available databases areto be created and maintained by an industry-sponsored entity recognized by the FCC andthe NTIA. They are to contain informationregarding existing and planned access BPLsystems. Each database should be availablewithin 30 days before initiation of the specificsystem’s service and should include thefollowing information:

– The name of the access BPL provider

– The frequency of the access BPLoperation

– The postal ZIP codes served by thespecific access BPL operation

– The manufacturer and type of access BPL equipment and its associated FCC identification, etc.

– Complete contact information for aperson at the BPL operator’s company in charge of resolving any interferencecomplaints

– The proposed or actual date of accessBPL operation

• Interference Mitigation and Avoidance:Access BPL systems are basically required toadhere to the NTIA recommendations forinterference mitigation and avoidancementioned above.

• Field Limits: Access BPL systems thatoperate in the 1.705-to-30-MHz band overMV lines must comply with the radiationlimits for intentional radiators provided inSection 15.209. Systems operating in the

Because BPL is a new technology,

the FCC hasrequired that

all BPL-relatedequipment

be certified.Certification is an equipmentauthorization

by the FCC or itsdesignated entities,

as opposed toverification, which is

a manufacturer’sself-approval

procedure.


30-to-80-MHz band over MV lines mustcomply with the radiation limits forunintentional radiators provided in Section15.109 (b). Systems operating over LV linesmust comply with the Section 15.109 (a) and(e) limits. Radiation emission limits for accessBPL equipment are summarized in Table 1.

The FCC also decided to eliminate conductedemission limits and testing for BPL systemsbecause of the danger and inconvenienceassociated with measuring power lineconducted emissions.

• Measurement Procedure and Guidelines:The FCC requires that access BPL systememissions be measured in situ to demonstratecompliance with the new Part 15 rules.Measurements are to be made at a minimumof three overhead and three undergroundrepresentative points and according to themeasurement guidelines outlined inAppendix C of the NPRM. For access BPLsystems installed on overhead power lines, totake into account the effect of line length, thereceived measurement antenna will bemoved down-line parallel to the power line,starting from the access BPL signal injectionequipment location, to find the maximum

emissions at each frequency within thefrequency range of the access BPL device.The distance from the measurement antennato power line is the slant distance or range, as shown in Figure 5.

Because the distances r specified in the guidelinesmay coincide with unsafe locations (e.g., themiddle of a highway), the guidelines also specifyhow to extrapolate a distance correction factorfrom measurements made at distances other thanas specified in the rules. For frequencies below 30 MHz, the measured values are reduced by 40 log(10) (30/r); for frequencies at or above 30 MHz, the measured value is increased by 20 log(10) (r/10). The guidelines also specify thetype of measurement antenna (loop or linear) and the type of detector (peak, quasi-peak, or rootmean square [RMS]).

It is worth mentioning again that the FCCrecognized the interference potential of BPLsystems. That is why the FCC decided that, eventhough access BPL systems remain under thenewly added Subpart G of Part 15 for unlicenseddevice rules, their operations cannot causeharmful interference and the systems must acceptany outside interference. Furthermore, any BPLresultant interference must be corrected andresolved by the BPL operator immediately,without ceasing broadband service to the public.

On November 3, 2006, the FCC also decided toclassify BPL-enabled Internet access services asinformation services. By virtue of being consideredinformation services, BPL services become freefrom many, if not all, common carrier regulationsand associated fees and taxes. Specifically, the FCC’s Order finds that the transmission

The FCC recognizedthe interference

potential of BPL systems.

That is why the FCC decided that,

even though accessBPL systems remain

under the newlyadded Subpart G

of Part 15 for unlicenseddevice rules,

their operationscannot cause

harmful interferenceand the systems

must accept any outside

interference.

Power LineType

Frequency(MHz)

Field StrengthLimits (μV/m)

MeasurementDistance (m)

LV or MV 1.705–30 30 30

LV 30–80 100 3

MV 30–80 90 10

Table 1. Radiation Limits

RingAntenna

Antenna Height

Distance Specified in Rule (e.g., 30 m for <30 MHz)

Slant Range

Extrapolate at 2

0 or 40 log(R)

Figure 5. Interference Measurement Setup


component underlying BPL-enabled accessservices is telecommunications and thatproviding this telecommunications transmissioncomponent as part of a functionally integrated,finished BPL-enabled Internet access serviceoffering is an information service. The FCC’sdecision was based on its desire to regulatesimilar services in a similar manner. The FCC’sOrder places BPL-enabled Internet access serviceson an equal regulatory footing with otherbroadband services such as DSL or cable modemInternet access services [34].

The FCC may, however, still decide to requireBPL operators who provide VoIP services tocontribute to the Universal Service Fund (USF),based on a percentage of their gross revenues.The USF was created by the FCC in 1997,following enactment of the TelecommunicationsAct of 1996, primarily to ensure that rural and low-income customers receive levels oftelecommunications service similar to those innonrural areas. All telecommunications carriersthat provide service internationally and between states are required to contribute to theUSF. The Universal Service AdministrativeCompany (USAC) submits fund size andadministrative cost projections for each quarter in accordance with FCC rules.

The FCC also released a new R&O in May 2006regarding law enforcement and emergencyservices [35]. More specifically, the FCC resolveda second R&O in the Communications Assistancefor Law Enforcement Act (CALEA) andBroadband Access Services proceedings. As aresult of this FCC Order, VoIP- and facilities-based broadband access providers, such as BPL operators who provide VoIP services, must bring their networks into compliance with wiretap, surveillance, and other official law enforcement and emergency servicesrequirements by May 14, 2007.

BUSINESS MODELS AND ECONOMIC ISSUES

Depending on their particular business and financial objectives, electric utility

companies can choose one of three businessmodels with respect to their BPL deployment. Aspresented below, each model has successivelymore associated risks and rewards:

• The Landlord or Retail Model: In thismodel, the electric utility company leases its facilities to another company (preferablyone with prior communications experience) that builds and operates the BPL system.

End users interface only with this companyfor all customer care, billing, and support.The electric utility company only collectsleases on its facilities, and may also receivesmart-grid services from the same BPL service builder/provider. This modelrequires the lowest investment from theelectric company and provides it with a newsource of income along with its existinginvestments. This is the lowest risk, if any,model for the electric company.

• The Wholesale Model: In this model, the electric company builds out the BPLnetwork and leases it to another company,which wholesales the bandwidth tocommunications service providers orInternet service providers (ISPs) that operatethe network and interface with customers.This is a medium risk option, and the BPLnetwork can be used to provide smart-gridservices for the electric company.

• The Service Provider Model: This is themost aggressive model. The electric utilitycompany builds and operates the BPLnetwork and interfaces directly with thecustomers. Here, the electric company needsto acquire the communications expertiserequired to build, operate, and maintain the BPL network. Of course, the electriccompany must also market the broadbandservices. This model carries the most risk, but offers the greatest potential return oninvestment (ROI).

Currently, precise data regarding BPLdeployment costs is not publicly available.Various estimates show that BPL costs per homepassed could range from $50 to $300, dependingon the electric grid’s architecture, the need forrepeaters, the number of homes connected to thesubstation, and similar factors. This cost includesnot only the cost of equipment and installation,but also the cost over time of maintenance,equipment replacement, and upgrades.Consumer premises equipment (CPE) costscurrently range from $50 to $200. Assuming aconservative initial deployment with a subscriberpenetration rate of 10 percent (blended over rural,suburban, and urban areas), which is typical of current initial deployment results, and a $100-per-home-passed deployment cost and a$100 CPE cost, the initial BPL deployment costbecomes about $1,100 per subscriber. This number is in line with numbers published in the final BPL report from United Telecom Council (UTC) Research and The Shpigler Group,

In November 2006,the FCC decided

to classify BPL-enabled

Internet accessservices asinformation

services. By virtue of

being consideredinformation

services, BPLservices becomefree from many,

if not all, common carrierregulations andassociated fees

and taxes.


which compares deployment costs for variousbroadband technologies [36, 37]. See Figure 6.

It is also interesting to note that, even thoughdeploying BPL in rural areas could be lessexpensive than deploying DSL, cable, or fiber, itmay still be prohibitively expensive per capita.With this in mind, BPL operators may choose,instead, to compete with DSL, cable, and otherservice providers in suburban and urban areaswhere some sort of broadband services alreadyexists. Ironically, this would defeat the mainreason that the FCC adopted BPL: to acceleratethe availability of broadband services inunderserved areas. Furthermore, prior experienceand research have shown that BPL service needs to be either significantly better (e.g., havehigher user throughputs), cheaper, or both, to be able to convince subscribers to change existing services to BPL or to attract newsubscribers to this new technology.

With this in mind, BPL service penetration PBPLwould typically be some function of BPL servicecost CBPL , including CPE, installation and setup,and a monthly service fee; the service costs ofexisting broadband services Cexisting ; the availabledata throughput of BPL RBPL ; and the datathroughput of existing services Rexisting [38]. Asimple formulation could be:

where α is a weighting factor (e.g., 10 or 20) that reflects the importance of performanceversus cost.

In this formulation, PBPL becomes null if its costand data rates are the same as those of existingbroadband services. Of course, this formulationdoes not take into account the value that BPLoffers by providing smart-grid services.(Assessing the potential revenue and savingsfrom BPL smart grid services would be thesubject of another study.)

CONCLUSIONS

Even though the importance and directsocioeconomic impact of access to broadband

services are well understood, currently only 4 percent of the Earth’s population has access to some type of broadband services, typically via DSL or cable modem. BPL offers a new,potentially powerful alternative means ofproviding high-speed Internet services, VoIP, andother broadband services to homes andbusinesses by using existing MV and LV powerlines. Because roughly 60 percent of Earth’sinhabitants have access to power lines, BPL couldplay a significant role in bridging the existingdigital divide. But the success of BPL, like that of any new technology in its infancy,depends on more than strong theoretical results or successful field testing. It also dependsgreatly on the appropriate business models anddeployment plans.

As the regulatory uncertainties and interferenceissues surrounding BPL dissipate, and with thesuccess of many field trials and early commercialdeployments, the release of various standards,and the growing availability of reasonably pricedstandardized and reliable equipment, the road to BPL is becoming increasingly well paved and broadband over power lines seems to be well energized. Indeed, BPL’s future looks very bright!

ACKNOWLEDGMENTS

One of the authors, S. Rasoul Safavian, would like to express his gratitude for useful

discussions with Professor Mohsen Kavehrad ofthe Electrical Engineering Department at thePennsylvania State University, several staffmembers of the Federal CommunicationsCommission, and David Shpigler of The Shpigler Group.

Acce

ss M

etho

d

Wireless

DSL

Cable Modem

BPL

Satellite

FTTH

Deployment Cost per Subscriber ($)

$1,825

$1,408

$1,007

$900

$828

$800

0 400 800 1,200 1,600 2,000

Figure 6. Deployment Costs for Different Access Technologies

PBPL = Min{100, Max{0, [(Cexisting–CBPL)

+α [log2(RBPL) – log2( Rexisting)]]}}

As the regulatoryuncertainties and

interference issuessurrounding

BPL dissipate, andwith the success of

many field trials andearly commercial

deployments, the release of

various standards,and the growing

availability ofreasonably pricedstandardized and

reliable equipment,the road to BPL

is becomingincreasingly well paved.


TRADEMARKS

Amperion is a trademark of Amperion, Inc.

CURRENT Technologies is a registeredtrademark of CURRENT CommunicationsGroup, LLC.

EarthLink is a registered trademark of EarthLink,Inc.

HomePlug is a registered trademark of theHomePlug Powerline Alliance.

P1675 is a trademark of the IEEE.

REFERENCES

[1] T. Farrar, “ATC: The Future of Mobile SatelliteServices?” TMF Associates report, publishedFebruary 2006/updated August 2006(http://www.tmfassociates.com/ATC.html).

[2] Information about and images of the Earthin a website by The Ozone Hole, Inc.(http://www.solcomhouse.com/earth.html).

[3] C. Howson, “Technology Outlook for Broadband & Mobile Networks,” BROADWANWorkshop, Cuenca, Spain, November 2005(http://www.telenor.no/broadwan/WorkshopNovember2005/Howson_BROADWAN_Cuenca_Workshop.pdf).

[4] T.L. Friedman, “The Lexus and The Olive Tree:Understanding Globalization,” Farrar, Straus and Giroux, 2000.

[5] “Promoting Innovation and Competitiveness –President Bush’s Technology Agenda”(http://www.whitehouse.gov/infocus/technology/).

[6] United Power Line Council (UPLC) BPLDeployments Map (http://www.uplc.utc.org/file_depot/0-10000000/0 10000/7966/conman/BPL+Map+12_12.pdf).

[7] Open PLC European Research Alliance (OPERA)(http://www.ist-opera.org/).

[8] IEEE BPL Study Group and Working Groupmeeting information re Standardization of Broadband Over Power Line Technologies(http://grouper.ieee.org/groups/bpl/index.html).

[9] Universal Powerline Association (UPA)(http://www.upaplc.com/page_viewer.asp?category=Home&sid=2).

[10] HomePlug Powerline Alliance(http://www.homeplug.org/en/index.asp).

[11] S. Houle, “Building the Smart Grid – TXU BPLInitiative” (http://www.gulfcoastpower.org/default/2-06meeting-houle-austin.pdf).

[12] J.M. Bradbury, “Broadband over Power Lines a Foundation for the Utility of the Future,”Energy Summit 2006 (http://www.enrg.lsu.edu/conferences/energysummit2006/Jay_Bradbury_2.pdf).

[13] M. Gotz, M. Rapp, and K. Dostert, “Power Line Channel Characteristics and Their Effect on Communication System Design,” IEEE Communications Magazine, Vol. 42, No. 4,April 2004, pp. 78-86.

[14] H. Ferreira, O. Hooijen, and H. Grove, “Power Line Communications: An Overview,”IEEE Conference Proceedings for AFRICON 1996,Stellenbosch, South Africa, September 1996, Vol. 2, pp. 558–563.

[15] M. Zimmermann and K. Dostert, “A Multipath Model for the Powerline Channel,”IEEE Transactions on Communications, Vol. 50, No. 4, April 2002, pp. 553–559.

[16] M. Zimmermann and K. Dostert, “Analysis andModeling of Impulsive Noise in Broad-bandPowerline Communications,” IEEE Transactions on Electromagnetic Compatibility, Vol. 44, No. 1,February 2002, pp. 249–258.

[17] J.R. Carson, “Wave Propagation in OverheadWires with Ground Return,” Bell Systems Technical Journal, Vol. 5, May 1926, pp. 539–554.

[18] M. D’Amore and M.S. Sarto, “A NewFormulation of Lossy Ground Return Parametersfor Transient Analysis of Multi-ConductorDissipative Lines,” IEEE Transactions on PowerDelivery, Vol. 12, No. 1, January 1997, pp. 303–314.

[19] P. Amirshahi and M. Kavehrad, “High-FrequencyCharacteristics of Overhead MulticonductorPower Lines for Broadband Communications,”IEEE Journal on Selected Areas in Communications,Vol. 24, No. 7, July 2006, pp. 1292–1303.

[20] P. Amirshahi and M. Kavehrad, “TransmissionChannel Model and Capacity of Overhead Multi-Conductor Medium-Voltage Power-Lines forBroadband Communications,” IEEE Proceedings of CCNC 2005, Las Vegas, NV, January 2005, pp. 354–358.

[21] J.J. Lee, “Measurements of the CommunicationsEnvironment in Medium Voltage PowerDistribution Lines for Wideband Power LineCommunications,” Proceedings of ISPLC 2004,Zaragoza, Spain, pp. 69–74.

[22] M. Zhang and W. Lauber, “Evaluation of theInterference Potential of PLC Systems,”Proceedings of the 2006 IEEE InternationalSymposium on Power Line Communications and its Applications (ISPLC-2006), Orlando, FL, March 2006, pp. 296–301.

[23] P.S. Henry, “Interference Characteristics ofBroadband Power Line Communication Systems Using Aerial Medium Voltage Wires,”IEEE Communications Magazine, Vol. 43, No. 4,April 2005, pp. 92–98.

[24] R.G. Olsen, “Technical Considerations forBroadband Powerline (BPL) Communication,”Proceedings of 15th International Zurich Symposiumon Electromagnetic Compatibility (EMC Zurich ‘05),Zurich, February 2005.

[25] L.S. Cohen, J.W. de Graaf, A. Light, and F. Sabath,“The Measurement of Broadband over PowerLine Emissions,” Proceedings of the 2005 IEEEInternational Symposium on ElectromagneticCompatibility (EMC 2005), Chicago, IL, Vol. 3, pp. 988–991.


[26] W. Liu, H. Widmer, and P. Raffin, “BroadbandPLC Access Systems and Field Deployment in European Power Line Networks,” IEEE Communications Magazine, Vol. 41, No. 5, May 2003, pp. 114–118.

[27] B. Malowanchuk, “Broadband over Power Line(BPL) Interference: Fact or Fiction?” La Revue desRadioamateurs Canadiens (Canada’s Amateur RadioMagazine), July & August 2004, pp. 39–44(http://www.arrl.org/tis/info/HTML/plc/files/Barry.pdf).

[28] H. Dai and H.V. Poor, “Advanced SignalProcessing for Power Line Communications,”IEEE Communications Magazine, Vol. 41, No. 5,May 2003, pp. 100–107.

[29] Corridor Systems, Inc.(http://www.corridor.biz/).

[30] ARRL comments on ET 03-104, the FCC Notice of Inquiry on Broadband Over Power Line(http://www.arrl.org/announce/regulatory/et03-104).

[31] NTIA Report 04-413, “Potential Interference fromBroadband over Power Line (BPL) Systems toFederal Government Radio Communications at 1.7 - 80 MHz – Phase 1 Study, Volume I,” NTIA, U.S. Department of Commerce, April 2004(http://www.ntia.doc.gov/ntiahome/fccfilings/2004/bpl/).

[32] FCC 04-29, Notice of Proposed Rule Making, in the Matter of Carrier Current Systems,including Broadband over Power Line Systemsand Amendment of Part 15 regarding newrequirements and measurement guidelines forAccess Broadband over Power Line Systems,February 2004 (http://hraunfoss.fcc.gov/edocs_public/attachmatch/FCC-04-29A1.pdf).

[33] FCC Report and Order 04-245, ET Dockets 04-37and 03-104, in the Matter of Amendment of Part 15 regarding new requirements andmeasurement guidelines for Access Broadbandover Power Line Systems and Carrier CurrentSystems, including Broadband over Power LineSystems, October 2004(http://www.hraunfoss.fcc.gov/edocs_public/attachmatch/FCC-04-245A1.doc).

[34] W.D. Gardner, “FCC Endorses Broadband over Powerline,” TechWeb Technology News,November 3, 2006 (http://www.techweb.com/wire/mobile/193501695).

[35] FCC 06-56, Second Report and Order andMemorandum Opinion and Order, ET Docket No. 04-295 and RM-10865, in the Matter ofCommunications Assistance for Law EnforcementAct and Broadband Access and Services, released May 2006/finalized December 2006.

[36] “Opportunities in Broadband over Power Line,”Report by The Shpigler Group and UTC Research,July 2004 (http://www.igigroup.com/st/pages/utc.bopl.html).

[37] David Shpigler, private communication,December 2006.

[38] P.A. Brown, “Identifying Some Techno-Economic Criteria in PLC/BPL Applications and Commercialization,” Proceedings of the 9th International Symposium on Power LineCommunications (ISPLC) and its Applications, Vancouver, Canada, April 2005, pp. 234–239.

BIOGRAPHIESLee Lushbaugh, principal vicepresident, Bechtel Corporation,and general manager, Tele-communications, Americas,provides day-to-day oversightfor both business developmentand operational activities in the region. During 2006, the regional staff reachedapproximately 1,500 employees

working in 35 markets across the continental UnitedStates. Previously, Lee has served as director ofengineering and as the program director of severalnationwide wireless programs and a fiber deploymentprogram. He joined Bechtel Telecommunications in1996 as vice president/manager of engineering and wasthe initial developer of its engineering department. Lee joined Bechtel Corporation in 1974 and, beforejoining Bechtel Telecommunications, held bothfunctional and operational roles in the fossil power andnuclear business lines, including the plant design, civil,and mechanical engineering disciplines. Lee received a BS in Mechanical Engineering from the University of Maryland. He is a RegisteredProfessional Engineer in various states, a member of theAmerican Society of Mechanical Engineers, and a SixSigma Champion.

Rasoul Safavian brings morethan 15 years of experience in the wired and wirelesscommunications industry to his position as BechtelTelecommunications’ vicepresident of Technology,Americas Regional BusinessUnit. He is charged withestablishing and maintaining

the overall technical vision for Bechtel’s Americanmarkets and providing guidance and direction to its specific technological activities. In fulfilling this responsibility, he is well served by his background in cellular/PCS, fixed microwave, satellitecommunications, wireless local loops, and fixednetworks; his working experience with major 2G, 2.5G,3G, and 4G technologies; his exposure to the leadingfacets of technology development as well as itsfinancial, business, and risk factors; and his extensiveacademic, teaching, and research experience.Before joining Bechtel in June 2005, Dr. Safavianoversaw advanced technology research anddevelopment activities, first as vice president of theAdvanced Technology Group at Wireless Facilities, Inc.,then as chief technical officer and vice president ofengineering at GCB Services. Earlier, over an 8-yearperiod at LCC International, Inc., he progressedthrough several positions. Initially, as principalengineer at LCC’s Wireless Institute, he was in charge of CDMA-related programs and activities. Next, as lead systems engineer/senior principal engineer,


he provided nationwide technical guidance for LCC’sXM satellite radio project. Then, as senior technicalmanager/senior consultant, he assisted key clients withthe design, deployment, optimization, and operation of3G wireless networks.Dr. Safavian has spoken at numerous conferences and industry events and has been publishedextensively, including technical papers in the previous three issues of the Bechtel TelecommunicationsTechnical Journal.Dr. Safavian is quite familiar with the ElectricalEngineering departments of four universities: TheGeorge Washington University, where he has been anadjunct professor for several years; The PennsylvaniaState University, where he is an affiliated facultymember; Purdue University, where he received his PhD in Electrical Engineering, was a graduate researchassistant, and was later a member of the visiting faculty;and the University of Kansas, where he received bothhis BS and MS degrees in Electrical Engineering andwas a teaching and a research assistant. He is a seniormember of the IEEE and a past official reviewer ofvarious transactions and journals.


INTRODUCTION

Called “the next big thing in software,”service-oriented architecture (SOA) promises

to change the way enterprises do business,reducing the cost and complexity associated withintegrating software applications. The primarydriver for change revolves around integration—specifically, a methodology for ease ofcommunications among applications. Asenterprises scale, the number of applicationsrequired to accomplish basic business processesscales as well; to accomplish a business function,the need to share data among applicationsbecomes apparent. Getting applications to talk toone another is a problem that has plaguedinformation services departments for decades.Cottage industries have been built around thisissue, with millions of hours in productivity andhuge dollar amounts in services and softwarespent trying to address it. Building on theinnovations used to create the Internet’scommunications structure, SOA promises tochange the approach to enterprise integration,potentially eliminating the headaches involved inthe inter-working of applications [1].

The framework for SOA extends well beyondenterprise computing; in fact, the SOA model canbe applied in telecommunications providernetworks to both increase the speed of launch fornew services and reduce the required networkinfrastructure.

BACKGROUND

Fundamental to SOA is the loose integration ofservices and applications. What does this

mean? To better understand loose integration asit pertains to an enterprise network, a basicunderstanding of distributed computing and tightintegration of applications is helpful.

As enterprises began to automate theiroperations, multiple computing platforms wereintroduced. For example, an enterprise may haveinitially purchased a mainframe and usedspecialized software to perform accounting andfinancial tasks. That same enterprise may haveused other software and systems to performscientific computing, graphic design, documentpublishing, payroll processing, inventory, etc. Atsome point, the enterprise recognizes the benefitsof integrating the software platforms. Consider,for instance, a business transaction involving theenterprise’s purchasing system and accountingsystem, each running on a different computingplatform. When a bill is generated by thepurchasing system, the outstanding amountneeds to be passed to the accounting system as areceivable item. But first, the message must betranslated, since the two platforms use twodifferently written software packages. A separatepiece of software is written to convert theinformation flowing from the purchasing systeminto data readable by the accounting system

SERVICE-ORIENTED ARCHITECTURE

Abstract—Service-oriented architecture creates a framework wherein applications can use standardized Webservices to share data. Using standardized Web services eliminates the need for proprietary, custom-developedmiddleware to specifically address how particular applications speak to one another. Eliminating middlewaresignificantly reduces the cost and complexity of enterprise networks. Telecommunications service providers canbuild on the gains realized in enterprise architectures. Carriers benefit from deploying these architectures, bothin managing their networks and customers and in deploying new services and applications.

Key Words—application integration, distributed architecture, IMS, OSS/BSS, service-oriented architecture,SOA, Web services


Brian [email protected]

(see Figure 1). An additional process within thiscustom interface software, or a separate piece ofsoftware, is also required to convert the

information flowing in the opposite direction.This deployment scenario is known as tightintegration of applications [2].

For tightly integrated networks, developingcustom interface software can prove to be achallenge. Developers must understand the innerworkings of both applications; extensive testing isalso required to ensure that all message and dataformats can be passed successfully. Datastructures in commercial software packages aretightly controlled as part of the intellectualproperty of the developer, further adding to the challenge. Network operations andmaintenance in this tight integration scenario canalso be troubling. Each application needing



AS accounting system

BGCF breakout gateway control function

BSC base station controller

BSS billing support system

BTS base transceiver station

CORBA® Common Object Request Broker Architecture

CSCF call session control function

ETSI European TelecommunicationsStandardization Institute

GGSN gateway GPRS support node

GPRS general packet radio service

HLR home location register

HSS home subscriber service

HTTP hypertext transport protocol

I-CSCF interrogating CSCF

IMS IP multimedia subsystem

IOOP inter-ORB object protocol

IP Internet Protocol

IPSec IP security

ITU-T International Telecommunication Union-Telecommunication Standardization Sector

MGCF media gateway control function

MGW media gateway

An enterprise mayface the choice ofwaiting to roll outnew software or

spending thousandsof dollars to updateinterface software toaddress changes in

a new release.

PurchasingSystem

CustomInterface

AccountingSystem

Figure 1. Separately Developed Custom Interface AllowsApplications To Pass Information

MRF multimedia resource function

MRFC MRF control

MRFP MRF processor

ORB object request broker

OSS operation support system

P-CSCF proxy CSCF

PDF policy decision function

PLMN public land mobile network

PSTN public switched telephone network

RAN radio access network

RMI remote method invocation

RNC radio network controller

SCIM service capability interaction manager

S-CSCF serving CSCF

SGSN serving GPRS support node

SIP session initiation protocol

SOA service-oriented architecture

SOAP simple object access protocol

UDDI Universal Description, Discovery, and Integration

WLAN wireless local area network

W3C® World Wide Web Consortium

WSDL Web services description language

XML extensible markup language

communication requires a separate piece of software to be executed and maintained,adding to software and hardware cost, as seen inFigure 2. Each new release of software mustundergo detailed evaluation to ensure that anychanges to data structures are understood. Forany changes identified, the software interfacesmust be modified; the process of executing,testing, and troubleshooting issues slows thedeployment of new software and adds cost forthe enterprise. An enterprise may face the choiceof waiting to roll out new software or spendingthousands of dollars to update interface softwareto address changes in a new release [3].

The complexity of tightly integrated applicationsonly increases as more applications are added.For example, turning to the business transactiondiscussed earlier, the enterprise may wish tointroduce a third software application, such as aninventory system, into the existing purchasingand accounting systems. Integrating thisapplication with both existing systems willrequire two additional software interfaces to bespecified, developed, deployed, and maintained.A firm can quickly expend a great amount of both financial and human resources ondeveloping and maintaining all of these softwareinterfaces [4].

Fortunately for those enterprises seeking to useand leverage the model of integrated softwareapplications, the development of the Internet andthe World Wide Web has led to better solutionswith the emergence of SOA. Using a conceptknown as Web services, SOA is a model for looseintegration of disparate applications.

SOA EXPLAINED

SOA focuses on the perspective that softwareapplications are services that support a

particular business process. Fundamental to theconcept of SOA are standards that determine howthese services are built, maintained, and used inthe network. By defining a standard andrequiring adherence to it, SOA helps to addressthe issues outlined earlier in describing a tightlyintegrated network.

The SOA concept revolves primarily around theuse of Web services to facilitate communicationsamong applications. Web services generally use alanguage based on extensible markup language(XML) to both standardize data formats andexchange information. The language is referred toas Web services description language (WSDL).Both XML and WSDL are World Wide WebConsortium (W3C®) standards, allowing the SOAframework to use data formats put in place byothers.

Using WSDL, system architects create a methodfor communications among applications that isfundamentally independent of the underlyingplatforms and programming languages. Themethod created is then advertised to the network as a “service,” a step referred to as“exposing” or “revealing” the application in anSOA framework [5].

Simple object access protocol (SOAP) is generallyused by computers to implement SOA. SOAPspecifies how a computer using hypertexttransport protocol (HTTP) and XML can executea program call to another computer; it alsospecifies how the program on the other computer


Figure 2. Integrating a New System Requires Further Custom Interface Development

Since Web services are platform-

independent andapplication-

transparent, even asweeping change,such as wholesale

replacement of an application,

can be done while remaininginvisible to the users and thearchitecture.

Purchasing System

Inventory System

CustomInterface

CustomInterface

Accounting System

CustomInterface


can format, package, and transmit a response tothe system calling it. HTTP and XML are presenton all computing platforms that use the Web,regardless of manufacturer, and SOAP takesadvantage of their presence. An added benefit isthat interoperability in the presence of firewalls isfurther assured by using HTTP and XML. While afirewall may be configured to block certain typesof traffic, based on port usage, protocols, orremote calls, almost all firewalls are configured toallow standard HTTP traffic. By using HTTP totransport XML, SOAP is most likely tosuccessfully penetrate corporate firewalls andother perimeter security devices. While SOAP isthe dominant methodology for communicationsover Web services, other techniques exist,including remote method invocation (RMI), using Java™; and inter-object request broker(ORB) object protocol (IOOP), a part of theCommon Object Request Broker Architecture(CORBA®) [6].

Once SOA is deployed in an enterprise,developers need only maintain the applicationsthat use the SOA, and not their method forcommunication. Again returning to the businessexample discussed earlier, if the purchasingsystem was to be updated, a developer wouldsimply make sure that the updates used Webservices properly for communication. Since Web services are platform-independent andapplication-transparent, even a sweeping change,such as wholesale replacement of an application,can be done while remaining invisible to the usersand the architecture, as long as the new platformcontinues to communicate with Web services [6].

The fundamental concepts of SOA have existedfor years, though described in different terms,such as “distributed architecture,” “modularprogramming,” and “event-oriented design.”Why did the SOA push prove successful, while

preceding concepts did not? With the rise of theInternet and the World Wide Web, almost everycomputer and operating system today ships withthe ability to access the Web using standardprotocols. Thus, the architects of the SOA conceptare able to leverage and build on these standardsto ensure interoperability among platforms, asseen in Figure 3.

Businesses can use WSDL to enter themselvesinto the Universal Description, Discovery, andIntegration (UDDI) registry, a list of businessesworldwide that use Web services. The UDDIregistry contains information on the business, aswell as details on how data is implemented in thefirm’s Web services. The registry streamlines thecommunications exchange among businesses,allowing applications from different businessesto interact and fostering e-commerce. Thus, SOAnot only improves integration within acompany’s network, it can also improveintegration among any number of companynetworks [7].

SOA AND TELECOMMUNICATIONS

As outlined, SOA can fundamentally changehow businesses integrate their applica-

tions and can streamline both enterprisecommunications and e-commerce. Can SOA also be used to improve the business of a telecommunications service provider? Thebusiness problems for service providers parallel those of large enterprises, and serviceproviders can benefit from the lessons learned by these enterprises.

A large area of focus and expense for a serviceprovider is the operation support system andbilling support system (OSS/BSS). A typicalOSS/BSS supports inventory management,provisioning, trouble tickets/resolution, lawfulintercept and compliance, customer relationshipmanagement, order management, and networkservices. Various applications accomplish one ormore of the many functions supported by theOSS/BSS, but there is no “one-size-fits-all,”single-supplier OSS/BSS implementation.Instead, organizations generally rely on ahodgepodge of vendor-provided and home-grown solutions, each performing part of afunction, a single function, or several. Serviceproviders also have to deal with legacy systems,some of which are built on foundations that are decades old. Integration of these systems is a great source of operational expenditures for carriers.

One of the greatestchallenges for

service providers inlaunching a new

application centerson supporting that

application;specifically,

how to provision,configure, manage,troubleshoot, and

bill for that service.

PurchasingSystem

AccountingSystemInventory

System

XMLInterface

Figure 3. Web-Based XML or WSDL Ensures Interoperability Among Platforms


Significant opportunities exist for integrating the applications that make up the OSS/BSS. Most of the data, such as subscriber informationor network infrastructure topology, is used bymultiple applications. To date, primarily point-to-point, proprietary systems have been developedand used to integrate the communications amongapplications. These systems are fraught with theissues outlined previously.

Carriers can work to expose each piece of the OSS/BSS with Web services and to build anSOA that links the pieces together, as shown inFigure 4. By doing so, a service provider can both extend the life of an existing system—particularly in the case of a homegrown or closed, proprietary solution—and simplify theintroduction of new systems and applications.Holistically, as the pieces of the OSS/BSS workbetter together, the subscriber experience shouldbe improved—including faster time-to-service and quicker, easier problem resolution. Thus, aservice provider can reduce customer churn bydeploying Web services to unify and integrate itssupport systems.

One of the greatest challenges for serviceproviders in launching a new application centerson supporting that application; specifically, how to provision, configure, manage,troubleshoot, and bill for that service. Withoutintegration, each piece of the OSS/BSS must beseparately addressed to ensure that the newservice functions properly—often resulting insignificant duplications of effort. Withintegration, applications can build on andleverage already defined frameworks from otherparts of the OSS/BSS. Service providers cantherefore reduce the time-to-market for newservices by deploying an SOA framework [8].

Standards bodies worldwide are exploring anddefining the benefits of using an SOA frameworkin network management. The EuropeanTelecommunications Standards Institute (ETSI)has released its architecture document TS 188 001,which outlines the architecture of a next-generation network’s operational supportsystem. The International TelecommunicationUnion-Telecommunication StandardizationSector (ITU-T) has produced RecommendationM.3060, “Principles for the Management of NextGeneration Networks.” Both of these documentsoutline how a network management systemfunctions and the benefits offered by SOA.

Consolidation is the norm in thetelecommunications industry; the seven BabyBells that resulted from the Modified FinalJudgment in 1983 have consolidated into threetoday. Mergers and acquisitions have also dottedthe landscape of cable system operators, wirelesscarriers, and competitive local exchange carriers.When a merger is executed, each firm brings apatchwork network of systems and operationalfunctions that must be integrated, along with thecustomers and networks of the firms involved.

Before SOA existed, a newly formed entity wouldhave to painstakingly inventory each system tounderstand its functionality, path forward, andinterfaces with other business processes beforedeciding to either (a) build a custom front end tointegrate the pieces, or (b) merge two platforms.Often, the same process was implementedmultiple times within each pre-merger entity,further complicating the integration project.

With SOA, Web services can expose multipleprocesses, allowing for seamless interaction ofdata among them. As processes are consolidated,the Web services used to communicate amongthem continue to function, removing theunderlying infrastructure from the userexperience. Deploying SOA allows the newlyformed firm to slowly and carefully integrate thepre-merger entity systems, easing transitions andremoving the urgency felt while evaluating thepath forward.

Verizon®, the nationwide local carrier formed bythe merger of the regional carrier Bell Atlanticand the independent consortium GTE, chose theSOA route when integrating the two networks.Verizon identified 500 key business functions andbuilt Web services around them. Known as the ITWorkbench, Verizon’s SOA even goes as far ashaving developers “advertise” the applicationsthey develop on a Web site, encouraging otherapplications to use them [9].

The extensibility ofapplications using

Web services makesthem ideal for

integration withthird parties; using

Web services, anetwork provider

can open itsnetwork to allow

third parties to offercontent and

applications.

Figure 4. Using SOA To Integrate OSS/BSS Systems

Inventory NetworkManagement

OrderManagement

OSS/BSSEnterprise SOA

Trouble-TicketTracking

Provisioning

Lawful Intercept


The SOA concept can go well beyond operationssupport for carriers; in fact, an SOA model can be used by a service provider in deployingnew applications and services. The extensibilityof applications using Web services makes themideal for integration with third parties; using Web services, a network provider can open itsnetwork to allow third parties to offer contentand applications.

Using SOA to leverage the network to providenew services turns the SOA model on its side;rather than just offering a way to save operationalcosts, SOA becomes a revenue generator for theoperator. The SOA model also streamlinesdevelopment time and costs for the third-partydeveloper. Standard information valuable to thedeveloper, such as the user’s handset type, is now readily available, saving the developer the time of writing the code to discover this typeof information.

T-Mobile® has taken this approach, sharinginformation about its network applications andresources via the SOA framework. Third partiesusing T-Mobile’s network can streamline thedevelopment of their applications, making use ofexisting infrastructure and services whereverpossible. T-Mobile provides external parties witha Web repository of deployed applications andservices, and how the services are implemented.This approach allows T-Mobile to collect higherproportions of dollars in revenue-sharingschemes with the third parties because the thirdparties’ costs to deploy applications are greatlyreduced by the SOA [10].

As wireless and other service providers begin todeploy integrated applications using an Internet

Protocol (IP) multimedia subsystem (IMS), asillustrated in Figure 5, comparisons are beingmade between the IMS and SOA. Is an IMS anSOA? As outlined earlier, SOA is not a particularproduct or standard, but a framework for a typeof architecture. Much like other types of SOAs, anIMS provides standardized, reusable services thatare separate from the underlying network;however, the IMS adds the dimensions of callsetup and control (via session initiation protocol[SIP]) and security (via IPSec). Security itself is not specified in the SOA framework; however, operators must consider security as afundamental piece of their architecture [11].

Thus, an IMS can be viewed as a special case ofSOA, designed in particular to supporttelecommunications applications and services.The IMS may also become just a portion of theSOA in an evolved network—another piece of theunderlying architecture used to deploy servicesto end users.

CONCLUSIONS

Thanks to the innovations of the Internet and the World Wide Web, a new

framework—SOA—has emerged that allows forthe integration of applications residing onvarious platforms and networks and even withinvarious firms. This framework holds greatpotential that is only beginning to be realized.Service providers can follow the example ofenterprises in deploying SOA to reduceoperational expenses and enhance the integrationof business processes. In addition, serviceproviders can use SOA principles to acceleratenew service offerings and increase revenues.Service providers are just beginning to realize the benefits of SOA, and the future looks bright for this architecture in the field of telecommunications.

TRADEMARKS

CORBA is a registered trademark of ObjectManagement Group, Inc., in the United Statesand/or other countries.

Java is a trademark of Sun Microsystems, Inc., in the United States and other countries.

T-Mobile is a federally registered trademark ofDeutsche Telekom AG.

The Verizon name is a registered trademark ofVerizon Trademark Services LLC or its affiliatesin the United States and/or other countries.

Service providerscan follow the

example ofenterprises in

deploying SOA toreduce operational

expenses andenhance the

integration ofbusiness processes.

ServicePlane

AS

S-CSCF

I-CSCF

P-CSCF

AS AS

SCIM

ControlPlane

MRFC

MRFP

PDF

HSS

HLRBGCF

MGCF

MGW MGWNode

B RNC

BTS

WLAN

BSC SGSN GGSNSGSN

Transport PlaneRAN

PSTNPLMN

Intranet/Internet

Figure 5. IMS Presents an SOA Framework


W3C is a registered trademark (registered in numerous countries) of the World Wide Web Consortium; marks of W3C are registeredand held by its host institutions MIT, ERCIM, and Keio.

REFERENCES

[1] “Service-Oriented Architecture (SOA)”(http://www-306.ibm.com/software/solutions/soa/overview.html?S_TACT=106AJ04W&S_CMP=campaign).

[2] E. Pulier and H. Taylor, Understanding EnterpriseSOA, Manning Publications Co., 2005.

[3] T. Erl, Service-Oriented Architecture: Concepts,Technology, and Design, Prentice Hall, 2005.

[4] D. Linthicum, “Application Integration:Addressing the Issues” (http://webservices.sys-con.com/read/44011.htm).

[5] “Will SOAP Wash Away Those BusinessIntegration Issues?” (http://searchwebservices.techtarget.com/originalContent/0,289142,sid26_gci883699,00.html).

[6] “Publishing and Finding Web Services UsingUDDI” (http://edocs.bea.com/wls/docs92/webserv/uddi.html).

[7] A. Skonnard, “SOAP: The Simple Object Access Protocol,” MSDN site(http://www.microsoft.com/mind/0100/soap/soap.asp).

[8] “OSS/BSS Integration” (http://www.capeclear.com/solutions/ossbss.shtml).

[9] L. Erlanger, “Verizon Goes Back to theWorkbench,” InfoWorld magazine(http://www.infoworld.com/article/05/11/07/45FEsoacaseverizon_1.html).

[10] C. Koch, “How SOA Really Works,” CIO magazine (http://www.cio.com/blog_view.html?CID=10591).

[11] C. Boulton, “When SOAs and Telcos Collide,”Enterprise magazine(http://www.internetnews.com/ent-news/article.php/3606016).

BIOGRAPHYBrian Coombe joined BechtelTelecommunications in 2003.Currently, as program managerof the Strategic InfrastructureGroup, a pivotal unit of theBechtel Federal Telecomsorganization, Brian manages a program that involvestelecommunications systemsand critical infrastructure

modeling, simulation, analysis, and testing. Heevaluates government telecommunications markets,formulates requirements for telecommunications andwater infrastructure work, and develops the StrategicInfrastructure Group’s scope.

As Bechtel’s technical lead for all radio frequencyissues, Brian draws on his extensive knowledge ofwireless and fiber optic networks. In his first positionwith the company, he engineered configurations toallow for capacity expansion of the AT&T WirelessGSM network in New York as part of a nationwidebuildout contract. Later, he was the lead engineer forplanning, designing, and documenting a fiber-to-the-premises network serving more than 20,000 homes. Heis the Bechtel Telecommunications Laboratory’sresident expert for optical network planning,evaluation, and modeling.

Before joining Bechtel, Brian was a systems engineer atTellabs®, where he launched the company’s densewavelength-division multiplexing services andmanaged network design and testing. He developedsolutions to complex network issues involving echocancellation, optical networking, Ethernet, TCP/IP,transmission, and routing applications.

Brian is currently completing work toward an MS in Telecommunications Systems Engineering at the University of Maryland. He earned a BS with honors in Electrical Engineering at The PennsylvaniaState University.

Brian is a member of the Institute of Electrical andElectronics Engineers; INSA; AFCEA; and Eta KappaNu, the national electrical engineering honor society. He has published technical articles in the Bechtel Telecommunications Technical Journal, and histutorial on Micro-Electrometrical Systems and OpticalNetworking was presented by the InternationalEngineering Consortium.


INTRODUCTION AND HISTORICAL BACKGROUND

Serious research in the field of outdoorpositioning first began in the 1960s, when

several US government agencies, including theDepartment of Defense (DoD), NationalAeronautics and Space Administration (NASA),and Department of Transportation (DOT),expressed interest in developing systems forposition determination [1]. The result, known asthe Global Positioning System (GPS), is the mostpopular positioning system used today. Activityin this area has continued since cellular networksbegan to flourish in the 1990s, driven largely byregulatory requirements (such as E-911) forposition estimation.

While these developments were taking place,similar research and development (R&D) beganin the field of indoor positioning (also known aslocalization or location estimation), driven byemerging applications in the commercial, publicsafety, and military arenas. Commercialapplications range from tracking inventory in awarehouse to tracking children, the elderly, andpeople with special needs [2]. Location-sensitiveWeb-browsing and interactive tour guides formuseums are other examples [3]. In the publicsafety and military arenas, very accurate indoorpositioning is required to help emergencyworkers and military personnel effectively

respond to threatening situations and completetheir missions inside buildings. Accurate indoorlocalization is also an important part of various personal robotics applications [4] and ofcontext-aware computing [5]. More recently,location sensing has found applications inlocation-based handoffs in wireless networks [6],location-based ad hoc network routing [7], and location-based authentication and security.

Many of these applications require low-cost, low-power terminals that can be easily deployed withlittle or no planning; this is the basis fordevelopments in ad hoc sensor networks. Recentdevelopments in integrated circuit (IC)technology and micro-electromechanical systems(MEMSs) have made it possible to realize suchlow-cost, low-power terminals. As a result, in thenext few years, scores of new applications forindoor localization will undoubtedly emerge.

Unfortunately, positioning techniques developedfor GPS and cellular networks generally do notwork well indoors. A primary reason is the signalattenuation caused by building walls. In addition,indoor radio channels exhibit much strongermultipath characteristics than outdoor channels.Furthermore, the accuracy requirements forindoor positioning systems are typically muchhigher than those for outdoor systems. For anoutdoor application such as E-911, an accuracy

INVITED PAPER

TECHNICAL ASPECTS OF LOCALIZATION IN INDOOR WIRELESS NETWORKS

Abstract—Location-aware wireless networking is an exciting new area in telecommunications. It is well-known that knowledge of a user’s location enables a number of location-based services to be delivered to thatuser. While this issue has been largely addressed for the outdoor case (most notably in cellular networks),accurate indoor positioning (also known as indoor localization or indoor location estimation) is still an open areaof research. This paper discusses the technical aspects of indoor localization systems. It reviews the challengesimposed by the indoor environment, the techniques used to surmount those challenges, and the state of the artin accurate localization techniques, with an emphasis on the quality of estimation.

Key Words—indoor geolocation, indoor localization, indoor location estimation, wireless networks


Muzaffer Kanaan1, 2

[email protected]

Mohammad Heidari2

[email protected]

Ferit Ozan Akgül2

[email protected]

Professor KavehPahlavan, PhD2

[email protected]

1 Verizon Laboratories2 Center for Wireless

Information NetworkStudies (CWINS),Worcester PolytechnicInstitute

of 125 m 67 percent of the time is consideredacceptable [8], while a similar indoor applicationtypically requires an accuracy level on the orderof a few meters [9]. For all of these reasons, newmethods of position estimation need to bedeveloped for the indoor setting.

Another factor to be considered is that therequired degree of location estimation accuracy istypically application dependent. For example, inan application such as inventory tracking in alarge warehouse, estimation accuracy on the

order of a few meters might be acceptable. Incontrast, much higher accuracy is needed forlocation estimates used in a public safety ormilitary application (such as situationalawareness systems). This makes it critical tocharacterize the quality of estimation (QoE) of agiven location estimation system. QoE is afunction of the particular algorithm used toestimate location, the quality of the informationavailable to the algorithm, impairmentsintroduced by the channel, and suboptimalcoverage conditions that are an ever-presentreality in any wireless network.

The rest of this paper is organized into four majorparts. First, the signal characteristics (also knownas location metrics) used to estimate location are examined and the effects of channelconditions on them are discussed. Second, the effects of suboptimal coverage conditions on localization are described. Third, a non-exhaustive survey of localization techniques forthe indoor environment is presented. Fourth, QoE is assessed.

FUNDAMENTALS OF INDOOR LOCALIZATIONSYSTEMS

Structure of a Localization SystemThe basic structure of a positioning system isillustrated in Figure 1, which shows a sensorwhose location is to be determined. The system



AOA angle of arrival

AP access point

BW bandwidth

CDF cumulative distribution function

CMS coverage map search

CN-TOAG closest-neighbor with TOA grid

DME distance measurement error

DoD Department of Defense

DOT Department of Transportation

DP direct path

FDP first detected peak

GPS global positioning system

IC integrated circuit

LOS line of sight

Required estimation accuracy

in an indoorlocalization system

is applicationdependent.

PositioningAlgorithm

Sensor Position Estimate

RP-2RP-1

RP-3RP-NLocation Metrics from

the Network

Sensor

Location Metric Location Metric

Location MetricLocation Metric

Figure 1. General Structure of an Indoor Geolocation System

LS least-squares


MEMS micro-electromechanical system

MSE mean-square error

NASA National Aeronautics and Space Administration

PDF probability density function

QoE quality of estimation

R&D research and development

RP reference point

RSS received signal strength

TOA time of arrival

UDP undetected direct path

UWB ultrawideband

WLAN wireless local area network

consists of two parts: reference points (RPs) andthe positioning algorithm. The RPs are radiotransceivers whose locations are known withrespect to some coordinate system. Each RPmeasures various characteristics or locationmetrics of the signal received from the sensor.These location metrics are fed into the localizationalgorithm, which then estimates the location ofthe sensor.

Location metrics are of three main types:

• Angle of arrival (AOA)

• Time of arrival (TOA)

• Received signal strength (RSS)

Angle of Arrival As its name implies, AOA indicates the directionfrom which the received signal is coming. The RPs use special antenna arrays to estimatethe AOA. Figure 2 shows an example of AOA estimation in an ideal nonmultipathenvironment. The two RPs measure the AOAsfrom the sensor as 78.3 and 45.0 degrees,respectively. These measurements are used toform lines of position whose intersection is theposition estimate.

In real-world indoor environments, however,multipath effects generally result in AOAestimation error. This error can be expressed as:

(1)

where θtrue is the true AOA value, generallyobtained when the sensor is in the line-of-sight(LOS) path from the RP; θ^ represents theestimated AOA; and α is the AOA estimationerror. As a result of this error, the ability toestimate the sensor position is restricted to anarea defined by an angular spread of 2α, asillustrated in Figure 3 for the two-RP scenario.This clearly illustrates that to use AOA for preciseindoor localization, the sensor has to be in theLOS path from the RP, which is generally not thecase. Therefore, relying on AOA techniques aloneresults in a location estimate with a low accuracy,or low QoE. However, AOA techniques can beemployed in certain adverse channel conditionsto obtain more accurate TOA measurements, asdiscussed in the next section.

Time of Arrival In TOA-based localization systems, the TOA ofthe first detected peak (FDP) of the receivedsignal is used to determine the time of flight τ

and consequently the distance, between thetransmitter and the receiver. The basic conceptcan be illustrated with reference to the channelprofile shown in Figure 4. Because the speed oflight in free-space c is constant, the time of flight


Figure 2. Illustration of AOA

78.3°

45.0°

Sensor

RP-1

RP-2

α

Figure 3. Illustration of AOA in the Presence of Multipath

78.3°

45.0°

Sensor

RP-1

RP-2

θ ̂ = θtrue α

Figure 4. Illustration of Basic TOA Principles for Localization

Filtered Radio Channel and Ray Tracing Taps

Time (ns)

Ampl

itude

(mU)

0.12

0.10

0.08

0.06

0.04

0.02

050 100 150 200 250 300 350 400

Direct Path (Expected TOA)

First Detected Peak (Estimated TOA)

TOA Estimation Error (Ranging Error = 0.46 m)

Channel Profile (BW = 100 MHz)

Detection Threshold

±


along the direct path (DP) between thetransmitter and the receiver gives the truedistance between the transmitter and receiver as follows:

(2)

Figure 4 illustrates the channel impulse responsegenerated by ray tracing software for arbitrarytransmitter and receiver locations. The channelimpulse response is usually referred to as theinfinite-bandwidth (BW) channel profile becausethe receiver can detect every single detectablepath. In practice, BW is limited, and although the receiver detects the same set of paths, eachpath has a pulse shape. Adding these pulseshapes forms another signal, referred to as thechannel profile.

In a multipath environment, adding the pulseshapes from paths other than the DP causes thepeak of the channel profile to shift away from theexpected TOA, resulting in a TOA estimationerror. As a result, the channel introduces rangingerror (also referred to as distance measurementerror [DME] in the literature), given as:

(3)

where d^ is the estimated distance and d is the true distance.

There are two main sources of ranging error:multipath effects and undetected direct path

(UDP) conditions. Multipath effects result inreflected and transmitted paths being receivedalong with the DP. As system transmission BWincreases, the pulses become narrower and theTOA estimate nears the expected TOA, resultingin smaller ranging error [10]. More accurate TOAmeasurements result in a location estimate with ahigher QoE. The UDP condition occurs when theDP falls below the detection threshold of thereceiver, as shown in Figure 5 [11]. This generallyhappens at the edge of coverage areas, or whenlarge metallic objects are in the path between thetransmitter and the receiver. As a result, thedifference between the FDP and the DP is beyondthe dynamic range of the receiver, and the DPcannot be detected, as shown in Figure 5. UDPconditions typically result in much larger rangingerrors that can significantly degrade the QoE.Unlike multipath-based ranging error, UDP-based ranging error typically cannot be reducedby increasing the BW.

Ultrawideband (UWB) measurements in typicalindoor areas have shown that both multipath-based and UDP-based ranging errors follow aGaussian distribution, with mean and variancethat depend on BW [10]. The overall model can beexpressed as follows:

(4)

where Gmw ,σw and GmUDP,w ,σ UDP,w are theGaussian random variables (RVs) that refer tomultipath-based and UDP-based ranging error,respectively. The subscript w in both casesdenotes the BW dependence. The parameter ζ is abinary RV that denotes the presence or absence ofUDP conditions, with a probability densityfunction (PDF) given as:

(5)

where PUDP,w denotes the probability ofoccurrence of UDP-based ranging error.

UDP Condition ClassesTwo classes of unavoidable multipath UDPconditions occur in a typical indoor localizationscenario. The first class occurs when a largeobject such as an elevator or a metallic chamberblocks the DP between the transmitter and thereceiver. This class of UDP conditions is referredto as shadowed UDP because the huge metallicobject shadows the direct connection between

There are two sources

of ranging error:multipath and

UDP conditions.

d = c × τ

ε = d – d^

Figure 5. Illustration of UDP-Based DME

d = d + G(mw,σw) log(1+d)+ζ· G(mUDP,w ,σUDP,w)

^

f (ζ) = (1 – PUDP,w)δ (ζ – 1) +PUDP,wδ (ζ )

the transmitter and the receiver. In shadowedUDP conditions, the distance between thetransmitter and the receiver could be short, andthe total received signal power consequentlycould be large.

The second class of UDP conditions occurs inobstructed LOS environments with low receivedsignal power. Here, the DP power is below thedetection threshold because of the large distancebetween the transmitter and the receiver;however, other paths still arrive that have signalstrengths above the threshold level. This class isreferred to as natural UDP because it occursnaturally in any indoor area, even in the absenceof large metallic objects [12]. Shadowed UDPconditions are characterized by a substantial dropin DP power and large DMEs, while natural UDPconditions are characterized by a moderate dropin DP power and relatively smaller DMEs [13].

BW EffectsAs explained earlier, DME is caused either by limitations in system BW or by the occurrenceof UDP conditions. Figure 6 illustrates the effectof BW in a typical indoor TOA measurementscenario. As the BW is gradually decreased to 300,200, and 100 MHz (Figure 6a), the larger DMEvalues caused by the BW effects appear in theplots’ cumulative distribution function (CDF).Because DME caused by BW limitations can shiftthe peak of the detected first path of the channelprofile in either direction, negative DMEs arenow observable as well. BW reduction spreadsthe error range. For example, at a 100 MHz BW,

the errors are between –5 and +10 meters.Therefore, as may be seen in Figure 6b, BWs on the order of 10 MHz, used by GPS, areinsufficient; instead, BWs on the order of severalhundred MHz are needed to reasonably protectagainst the extensive multipath characteristicsfound in indoor areas. For, say, a 200 MHz BW,the DME range is on the order of –3 to +7 meters,which is fairly comparable with the UDP errors ofup to 7 meters observed for infinite BW.

One approach to reducing BW requirementsbelow these values is to use super-resolutionalgorithms for post-processing, as described in References [14 and 15]. However, to reduceDMEs to values below those observed in UDPareas with infinite BW, fundamentally differentapproaches are required; these are examined inthe following subsection.

Diversity EffectsIt may be presupposed that using diversitytechniques (i.e., frequency, time, or spacediversity) may help to reduce DME in much the same way as using them would incommunications. However, as shown inReference [13], these techniques are highlyinefficient in reducing DME. Frequency diversitymay be helpful to a certain extent in combattingmultipath-based, but not UDP-based, DME.

Precise Localization in the Absence of DPAOA techniques can also be employed in theabsence of DP to help obtain more accurate TOAmeasurements. Each resolvable path arriving at


Diversity techniques

are generally not helpful

in combattingranging error.

Figure 6. CDF of Error Based on Bandwidth

1.0

0.9

0.8

0.7

0.6

0.5

0.4

0.3

0.2

0.1

0

1.0

0.9

0.8

0.7

0.6

0.5

0.4

0.3

0.2

0.1

0-5 0 5 10

Distance Distance

-40 -30 -20 -10 0 10 20 30 40

Infinite BWBW = 300 MHzBW = 200 MHzBW = 100 MHz

Infinite BWBW = 100 MHzBW = 50 MHzBW = 10 MHz

CDF of Error of the First Detected Path for Different BWs

Band

widt

h

Band

widt

h

(a) (b)


the receiver can be associated with a certainreflection scenario (called a path index) that has acertain number of reflections and transmissions.Additionally, each path has a certain signalstrength and TOA. These two quantities can beregarded as the traceable features of the path.

As the sensor moves along a given direction, itstraceable features exhibit differential changes, butthe reflection scenario stays basically the same.This concept is shown in Figure 7. As the sensormoves along the dotted line toward the end of thecorridor, the arriving path evolves in differentialsteps in accordance with the differential distance(dd) traveled by the sensor. In this example, thenumber of reflections and transmissions is eachtwo and does not change along this particularmovement. Adaptive antennas and beam steeringcan be used at the receiver to focus on thisparticular path and try to trace it along the dottedline, because it exhibits smooth TOA behavior.Thus, this additional information from theindirect path can be used to estimate the arrival ofthe DP when it cannot be detected. Only a certainoffset between the arrival time of the indirect pathand the DP needs to be considered.

Received Signal StrengthRSS is a simple metric that can be measured andreported by most wireless devices. For example,the medium access control (MAC) layer of theIEEE 802.11 wireless local area network (WLAN)standard provides RSS information from allactive access points (APs) in a quasi-periodicbeacon signal that can be used as a metric forlocalization [16]. RSS can be used in two ways forlocalization purposes, as discussed next.

If the RSS decays linearly with the log-distancebetween the transmitter and receiver, it ispossible to map an observed RSS value to a

distance from a transmitter and consequentlydetermine the user’s location by using distancesfrom three or more APs. In other words:

(6)

where α is the distance-power gradient, X is theshadow fading (a lognormal distributed randomvariable), Pr is the received power, and Pt is thetransmitted power. While simple, this methodyields a highly inaccurate estimate of indoordistances because instantaneous RSS inside abuilding varies over time, even at a fixed location;this is largely due to shadow fading andmultipath fading. If, on the other hand, the RSSvalue at a given indoor point is known, then thelocation can be estimated as the point where theexpected RSS value most closely approximatesthe observed RSS value. This is the essence of the pattern recognition approach to positionestimation, which is discussed in greater detail inthe following section.

COVERAGE EFFECTS

Suboptimal coverage conditions exist in justabout every type of wireless network. Just

as the performance of an indoor wirelesscommunication system can be affected bycoverage deficiencies (manifested in ways such asdegradation of voice quality or data throughput),an indoor localization system can be similarlyaffected if an adequate number of locationmetrics cannot be obtained. In fact, somealgorithms, such as those using least-squares (LS)techniques, cannot be applied if the number ofRPs seen by the user is less than three. Thisimplies that location estimation systems have tobe robust enough to operate with a lack ofinformation. This is the motivation behindalgorithms such as the coverage map search(CMS) algorithm discussed in the next section.

In addition, even if an adequate number of TOA,AOA, or RSS measurements can be obtained, thequality of the measurements can degradenoticeably if coverage quality declines. Forexample, empirical studies have shown that UDP conditions occur more frequently at theedges of the coverage area and can introduceconsiderable error into TOA-based rangemeasurements [17]. This, in turn, degradeslocation estimation performance.

Each path has a certain

signal strength and TOA. These are

regarded astraceable features

of the path.

Figure 7. Illustration of Different Reflection Scenarios Related to TOA Estimation

DP Blocked(UDP Condition)

RP

dd

Sensor

RSSd = 10log10 Pr = 10log10 Pt – 10α log10 d +X


LOCALIZATION ALGORITHMS

Position estimation techniques can be grouped(a) in terms of whether the sensing

infrastructure used to measure location metrics isdeployed in a fixed or an ad hoc manner, or (b) according to how the position computationsare performed. In centralized algorithms, alllocation metrics are sent to one central node,which then carries out the computations. Indistributed algorithms, the computational load forthe position calculations is spread over allnetwork nodes. This paper concentrates oncentralized algorithms; for discussions of distri-buted algorithms, the reader is referred to therelevant literature (including References [18] to[24] and associated references contained therein).

Two centralized algorithms for fixed positionestimation—the closest-neighbor with TOA grid(CN-TOAG) [25] and the CMS—are the subjectsof this section.

Closest-Neighbor with TOA Grid AlgorithmThe CN-TOAG algorithm takes advantage of thefact that for any given configuration of RPs, theexact value of the TOA is known that should beobserved at any given indoor point [25]. Anexample is shown in Figure 8, in which a wholearea has been divided into a sequence of gridpoints. Assuming that the position of the RPs isknown accurately, a vector of four range values(one from each RP) can be associated with eachpoint; this vector is known as the range signature.The collection of range signatures for an area isknown as a TOA grid. The vector of observedrange measurements is compared with the range

signature at each point. The error value, e(r), isdefined as:

(7)

where D is the vector of observed rangemeasurements and Z represents the rangesignature at a point r = (x, y)T. The estimatedlocation is the location with the minimum valueof e (r). It has been shown that the CN-TOAGalgorithm can yield more accurate locationestimates (i.e., higher QoE) than traditionalalgorithms such as LS, provided that the spacing(h in Figure 8) between grid points is smallenough [25].

As can be seen, CN-TOAG is an example of a pattern-recognition algorithm, wherebyobservations are compared with database valuesthat characterize the indoor area. The algorithmalso requires centralized computation to keeptrack of the TOA grid data structure and of thecomputations required for the location estimate.

Coverage Map Search AlgorithmThe CMS algorithm is conceptually based on theCN-TOAG algorithm [25]. The general systemscenario is as shown in Figure 9, in which aregular arrangement of RPs is assumed. Each RP

Figure 8. Illustration of CN-TOAG Algorithm Concepts

d_4 d_3

d_1

d_2

TOA Grid Point (x_i, y_ j)

RP-4 RP-3

RP-1 RP-2L

Xh

L

y

Figure 9. System Scenario Used To Develop the CMS Algorithm

d_4 d_3

d_1 d_2

RP-4 RP-3

RP-1 RP-2L

X

L

V

e(r) = e (x,y) =⏐⏐D– Z(r)⏐⏐=⏐⏐D– Z(x,y)⏐⏐


performs a TOA-based range measurement to auser to be located. Like the CN-TOAG algorithm,the CMS algorithm relies on a TOA grid.

In a realistic indoor environment, there arecoverage deficiencies or holes throughout thearea covered by the RPs, as depicted in Figure 10.This means that a valid range measurement fromevery RP cannot be guaranteed at every point. Forexample, with the simple scenario depicted inFigure 10, there are areas where only three orfewer range measurements are available. Thisessentially rules out the applicability of moreconventional algorithms, such as LS, whichrequire a minimum of three range measurementsto function properly. Figure 10 is a somewhatsimplified depiction of the coverage for thedifferent RPs; in reality, the coverage areas ofvarious RPs are not as tightly confined, due tofactors such as shadow fading. Nevertheless, thisview is useful for developing insight into theCMS algorithm.

The range measurement deficiencies follow aspecified pattern. For example, in Area 1 ofFigure 10, a user is only able to communicate with

RP-1. Therefore, the CN-TOAG algorithm isrefined by defining a so-called coverage signature C, which is an array of all RPs that can communicate with the user at a specific point. For example, at any point within Area 1, C = [1, –1, –1, –1]T, where –1 indicates that theuser cannot communicate with a particular RP. Inthe current example, this implies that the user canonly communicate with and obtain a rangemeasurement from RP-1. Similarly, for all pointsin Area 2, C = [1, 2, 3, 4]T, indicating that in thisarea the user can communicate with all fourreferences. It is possible to characterize an entirearea with a two-dimensional array of C vectorswhich, for the purposes of this discussion, iscalled a coverage map.

It is clear that observations of C can providevaluable information in locating a user; this is the idea behind the CMS algorithm. In essence,the algorithm exploits the knowledge about missing range measurements to narrow down the user’s location to a specific area and then toestimate it. In other words, the CMS algorithmoperates on the premise that the lack of informationis, in itself, information that can be exploited.

Figure 10. Illustration of Partial Coverage for the CMS Algorithm

RP-4 RP-3

RP-2RP-1

Area 2

Region Q_cArea 1

L

L

X

Coverage Bounds of RP-1Coverage Bounds of RP-2Coverage Bounds of RP-3Coverage Bounds of RP-4

x

y


Specifically, the algorithm works as described inthe following paragraphs.

For a given vector of range measurements, apattern is derived. For example, suppose a user is located at point X, as shown in Figure 10. Atthis point, the user can communicate only withRPs 1, 2, and 4. Therefore, the range measurementvector D = [d1, d2, –1, d4]T, where –1 refers to amissing range measurement from RP-3 caused bycoverage limitations. For the current example, theCMS algorithm translates this to an equivalentrepresentation in the C vector space as:

(8)

The algorithm then searches the coverage map fora region Qc that is a subset Q of all points withinthe area A where the coverage signatures matchCm. In other words:

(9)

In the current example, region Qc is as shown inFigure 10.

Based on the coordinates of the different pointswithin Qc and the coordinates of the APs, a rangesignature Z(x,y) can be computed for each point(x,y) based on purely geometrical considerations,just as in the CN-TOAG algorithm.

The range measurement vector D is thencompared with all range signatures to find thepoint r where Z most closely approximates D. Inessence, this is equivalent to minimizing thefollowing objective function, with the additionalcondition that the search for the minimum isconfined to the region Qc :

(10)

where dk is the range measurement from the k-thRP, N is the total number of RPs with which theuser can communicate, and (Xk,Yk) are thecoordinates of the k-th RP.

The primary advantage of the CMS algorithm isthat it can be used with any number of rangemeasurements, whereas other algorithms, such asLS, require a minimum of three rangemeasurements. The main characteristic of theCMS algorithm is that it requires a central

computing entity in the network to generate thecoverage map and perform the search for theminimum. Although the algorithm has beenpresented in terms of the simplified coveragescenario of Figure 10, it is readily applicable tomore realistic coverage scenarios if the coveragemap is generated using any accurate indoor radiopropagation model [26].

QoE ASSESSMENT

As mentioned, the indoor environmentexhibits strong multipath characteristics and

propagation is highly site-specific. Collectively,this is referred to as channel effects. Channel effects introduce errors into the metrics (such asTOA and RSS) used in the location estimationprocess. The accuracy of a location estimate alsodepends on how the RPs are situated with respectto the user; collectively, this is referred to asgeometry effects.

With these points in mind, it is critical to be ableto answer the following questions to effectivelyevaluate the performance of location estimationsystems:

• How can the effectiveness of a given systembe gauged across different types ofbuildings?

• What is the probability of obtaining a givenperformance from a given localizationalgorithm x in any building configuration?

• How accurate should metrics measurementsbe to guarantee a given location estimationaccuracy throughout a given area for acertain application?

These and similar questions require thecharacterization of a given localization system’sQoE. Although QoE can be defined in severalways, the definition adopted in this paper isestimation accuracy as given by mean-squareerror (MSE):

(11)

where Rest is the estimated location of the userand Ract is the actual location.

If the number of RPs used in a TOA-based systemis known, the CDF of the MSE can be used togauge performance across different buildingtypes. This concept, known as the MSE Profile[27], uses the fact that for a given number of RPs,different building configurations give rise todiffering numbers of UDP conditions.

e (x,y) = dk – (x – Xk)2 + (y – Yk)22

k=1

N

Σ ⎞⎟⎠⎞⎟⎠

Qc = { Q ⊂ A: C = Cm} ∀C

Cm = [1,2, –1,4]T

MSEpos = E{⏐⏐Rest – Ract⏐⏐ }2


Now suppose that there is a TOA-based systemover an indoor area of a known size, where thenumber of RPs and the location estimationalgorithm are given. Assume that the MSE cannotexceed a certain value, MSEmax , at all points in thearea. This requirement implies that the rangingerror itself needs to be below a certain maximumvalue. Based on these pieces of information, anupper bound can be calculated on the requiredranging error at all points. The results ofcalculating and plotting this bound over a 20 x 20 m area for four RPs and for five RPs areshown in Figure 11 and Figure 12, respectively.These figures also clearly depict the dependenceof the estimation accuracy on geometry effects.

CONCLUSIONS

This paper has presented a technical overviewof indoor localization (also known as indoor

location estimation) systems. Indoor locationestimation is a relatively new area of research thatwill be an important enabler of future location-aware indoor wireless networks. Before this canhappen, however, a number of importantproblems must be solved in terms of optimizingalgorithm performance and characterizing QoE.Once this is accomplished, the future holds manydifferent applications for indoor locationestimation technology.

Figure 12. Upper Bound on Range Measurement MSE over All Points in a 20 x 20 m Area for Five RPs

202015

1510 105 50 0y (m) x (m)

Variation of the Range Measurement MSEas a Function of Actual Location(Desired Positioning MSE = 0.1)

Rang

e Mea

sure

men

t MSE

1.251.201.151.101.051.000.950.900.85

20 m

20 m

d_4 d_3

d_2d_1d_5

RP-3RP-4

RP-1 RP-2

RP-5

y

x

QoE depends on

geometry effects.

Figure 11. Upper Bound on Range Measurement MSE over All Points in a 20 x 20 m Area for Four RPs

0.105

0.100

0.095

0.090

0.085

0.080

0.07520

1510

50 0

510

1520

y (m) x (m)

Rang

e Mea

sure

men

t MSE

Variation of Range Measurement MSEas a Function of Actual Location

(Maximum Desired Positioning Error = 0.1)y

x

d_4 d_3

d_2d_1

RP-3RP-4

RP-1 RP-2

20 m

20 m


REFERENCES

[1] E.D. Kaplan, ed., Understanding GPS: Principles and Applications, Artech House, Boston, MA, 1996.

[2] M.L. McKelvin, M.L. Williams, and N.M. Berry,“Integrated Radio Frequency Identification andWireless Sensor Network Architecture forAutomated Inventory Management and Tracking Applications,” Proceedings of the Richard Tapia Celebration of Diversity in ComputingConference 2005, TAPIA ‘05, Albuquerque, NM,October 2005.

[3] S.G.M. Koo, C. Rosenberg, H.-H. Chan, and Y.C. Lee, “Location-based E-campus WebServices: From Design to Deployment,"Proceedings of the First IEEE International Conferenceon Pervasive Computing and Communications,PerCom 2003, Fort Worth, TX, March 2003.

[4] P. Jensfelt, “Approaches to Mobile RobotLocalization in Indoor Environments,” PhDThesis, Royal Institute of Technology, Stockholm,Sweden, 2001.

[5] A. Ward, A. Jones, and A. Hopper, “A NewLocation Technique for the Active Office,” IEEEPersonal Communications Magazine, Vol. 4, No. 5,October 1997, pp. 42–47.

[6] K. Pahlavan, P. Krishnamurthy, A. Hatami, M. Ylianttila, J. Mäkelä, R. Pichna, and J. Vallström, “Handoff in Hybrid Mobile DataNetworks,” IEEE Personal CommunicationsMagazine, Vol. 7, No. 2, April 2000, pp. 34–47.

[7] Y. Ko and N.H. Vaidya, “Location-aided Routing(LAR) in Mobile ad hoc Networks,” Proceedings of ACM/IEEE International Conference on MobileComputing and Networking 1998, MOBICOM’98,Dallas, TX, October 1998.

[8] FCC Docket No. 94-102, “Revision of theCommission’s Rules to Insure Compatibility withEnhanced 911 Emergency Calling Systems,”Federal Communications Commission, TechnicalReport RM-8143, July 1996.

[9] A.H. Sayed, A. Tarighat, and N. Khajehnouri,“Network-Based Wireless Location: ChallengesFaced in Developing Techniques for AccurateWireless Location Information,” IEEE SignalProcessing Magazine, Vol. 22, No. 4, July 2005, pp. 24–40.

[10] B. Alavi and K. Pahlavan, “Indoor GeolocationDistance Error Modeling with UWB ChannelMeasurements,” Proceedings of the 16th AnnualIEEE International Symposium on Personal IndoorMobile Radio Communications, PIMRC 2005, Berlin, Germany, September 2005.

[11] K. Pahlavan, P. Krishnamurthy, and J. Beneat,“Wideband Radio Channel Modeling for IndoorGeolocation Applications,” IEEE CommunicationsMagazine, Vol. 36, No. 4, April 1998, pp. 60–65.

[12] M. Heidari and K. Pahlavan, “A Model forDynamic Behavior of Ranging Errors in TOA-based Indoor Geolocation Systems,”accepted for presentation at the IEEE VehicularTechnology Conference, VTC2006-Fall, Montreal, Canada, September 2006.

[13] K. Pahlavan, F. Akgul, M. Heidari, and A. Hatami, “Indoor Geolocation in the Absence of Direct Path,” accepted for publication in IEEE Wireless Communications Magazine.

[14] X. Li and K. Pahlavan, “Super-resolution TOA Estimation with Diversity for IndoorGeolocation,” IEEE Transactions on WirelessCommunications, Vol. 3, No. 1, January 2004, pp. 224–234.

[15] K. Pahlavan, X. Li, and J.-P. Mäkelä, “IndoorGeolocation Science and Technology,” IEEECommunications Magazine, Vol. 40, No. 2, February 2002, pp. 112–118.

[16] P. Bahl and V.N. Padmanabhan, “RADAR: An In-Building RF-based User Location andTracking System,” IEEE Conference Proceedings,Vol. 2, pp. 775–784, INFOCOM-2000, Tel Aviv,Israel, March 2000.

[17] B. Alavi and K. Pahlavan, “Studying the Effect ofBandwidth on Performance of UWB PositioningSystems,” Proceedings of the IEEE WirelessCommunications and Networking Conference, WCNC 2006, Las Vegas, NV, April 2006.

[18] L. Doherty, K.S.J. Pister, and L. El-Ghaoui,“Convex Position Estimation in SensorNetworks,” IEEE Conference Proceedings, Vol. 3,pp. 1655–1663, INFOCOM-2001, Anchorage, AK,April 2001.

[19] D. Niculescu and B. Nath, “Ad-Hoc PositioningSystem (APS),” IEEE Conference Proceedings,GLOBECOM-2001, San Antonio, TX, November 2001.

[20] D. Niculescu and B. Nath, “Ad-Hoc PositioningSystem (APS) Using AOA,” IEEE ConferenceProceedings, INFOCOM-2003, San Francisco, CA,March/April 2003.

[21] G. DiStefano, F. Graziosi, and F. Santucci,“Distributed Positioning Algorithm for Ad-HocNetworks,” Proceedings of the InternationalWorkshop on Ultra Wideband Systems (IWUWBS2003), Oulu, Finland, June 2003.

[22] B. Denis and N. Daniele, “NLOS Ranging ErrorMitigation in a Distributed Positioning Algorithm for Indoor UWB Ad-Hoc Networks,”International Workshop on Wireless Ad-HocNetworks, IWWAN’04, Oulu, Finland, May/June 2004.

[23] J. Hightower and G. Borriello, “Location Systemsfor Ubiquitous Computing,” IEEE ComputerMagazine, Vol. 34, No. 8, August 2001, pp. 57–66.

[24] D. Niculescu, “Positioning in Ad-Hoc SensorNetworks,” IEEE Network, Vol. 18, No. 4,July/August 2004, pp. 24–29.

[25] M. Kanaan and K. Pahlavan, “CN-TOAG: A NewAlgorithm for Indoor Geolocation,” Proceedings ofthe 15th IEEE International Symposium on Personal,Indoor and Mobile Radio Communications, Vol. 3,pp. 1906–1910, PIMRC 2004, Barcelona, Spain,September 2004.

[26] K. Pahlavan and A. Levesque, Wireless Information Networks, 2nd Edition, John Wiley and Sons, 2005.

[27] M. Kanaan, F.O. Akgul, B. Alavi, and K. Pahlavan, “Performance Benchmarking of TOA-based UWB Indoor Geolocation SystemsUsing MSE Profiling,” to appear in the Proceedings of the IEEE Vehicular TechnologyConference, VTC2006-Fall, Montreal, Canada, September 2006.


ADDITIONAL READING

Additional information sources used todevelop this paper include:

• A. Hatami, and K. Pahlavan, “ComparativeStatistical Analysis of Indoor Positioning UsingEmpirical Data and Indoor Radio Channel Models,”IEEE Consumer Communications and NetworkingConference, CCNC 2006, Las Vegas, NV, January 2006.

• B. Alavi and K. Pahlavan, “Studying the Effect ofBandwidth on Performance of UWB PositioningSystems,” Proceedings of the IEEE WirelessCommunications and Networking Conference, WCNC 2006, Las Vegas, NV, April 2006.

BIOGRAPHIESMuzaffer Kanaan is a PhDcandidate at the WorcesterPolytechnic Institute, Massa-chusetts, and a distinguishedmember of the technical staff atVerizon Laboratories (R&Ddivision of Verizon Corpora-tion), where he is responsible forleading various activities in thefield of wireless access network

architecture, design, and optimization. Previously, heled optical network design initiatives within thecompany, most notably lab evaluations leading toVerizon’s nationwide long-haul optical backbonenetwork build-out.

Mr. Kanaan received his BS degree from EasternMediterranean University, Famagusta, North Cyprus,and his MS degree from New Jersey Institute ofTechnology, Newark, both in Electrical Engineering. He holds one US patent and another patent is pending.Mr. Kanaan is a member of the IEEE and Eta Kappa Nu.

Mohammad Heidari receivedhis MSc degree in Communica-tion and Computer Networkingfrom Worcester PolytechnicInstitute, Massachusetts, bydesigning a real-time RF testbedfor performance evaluation ofWiFi localization systems. He is currently pursuing his PhD at the Center for Wireless

Information Network Studies (CWINS), where hisresearch is focused on statistical modeling of radiochannel dynamic spatial behavior for indoorgeolocation applications.

Ferit Ozan Akgül received aBSc degree from Middle EastTechnical University, Turkey,and an MS degree, from KocUniversity, Turkey, both inElectrical Engineering. He iscurrently a PhD candidate at theCenter for Wireless InformationNetwork Studies (CWINS),Worcester Polytechnic Institute,

Massachusetts, where he works as a research assistanton RF-based indoor geolocation and indoor channelcharacterization, with a particular focus on angle-of-arrival diversity.

Kaveh Pahlavan is a professorof Electrical and ComputerEngineering (ECE) andComputer Science (CS) andfounding director of the Centerfor Wireless InformationNetwork Studies (CWINS),Worcester Polytechnic Institute,Massachusetts. Previously, hewas a visiting professor at the

Telecommunication Laboratory and Center for Wireless Communications, University of Oulu, Finland.Dr. Pahlavan is the editor-in-chief of the InternationalJournal of Wireless Information Networks, an advisoryboard member of the IEEE Wireless Magazine, and anExecutive Committee member of the IEEE PIMRC. He has been an IEEE fellow since 1996 and was a Nokia fellow in 1999 and a Fulbright-Nokia scholar in2001. He has served as the general chair and organizerof many IEEE events and has contributed to numerousseminal technical and visionary publications regardingwireless office information networks, homenetworking, and indoor geolocation science andtechnology. Dr. Pahlavan is the principal author of “Wireless Information Networks” (Allen Levesque,co-author), John Wiley and Sons, 1995, 2nd Ed. 2005,and “Principles of Wireless Networks – A UnifiedApproach” (P. Krishnamurthy, co-author), PrenticeHall, 2002. Additional information regarding his work can be found at www.cwins.wpi.edu.

Dr. Pahlavan received a PhD from WorcesterPolytechnic Institute, Massachusetts, and an MS degree from the University of Tehran, Iran, both inElectrical Engineering.


INTRODUCTION

To recapture the benefits that digital signalprocessors (DSPs) have established in a

variety of radio frequency (RF) applicationdomains in recent decades, it is necessary tooperate on the optical field vector as opposed tothe optical energy. This approach enables digitalcoherent communications and interferometricsensing applications to be deployed. However, it also requires overcoming a myriad oftechnological and architectural challenges. Thispaper describes proposed solutions. Reference [1]discusses the key methods used by secure digitalcoherent free-space optical communications fortactical applications.

Figure 1 depicts an integrated homodyne receiverfabricated and tested by CeLight1. The receivedoptical signal is split into two (arbitrary)

orthogonal polarizations (H’ and V’; only one isshown for clarity) and each is mixed with a localoscillator in an optical 90° hybrid. The hybridaccepts the signal S and the local oscillator L andproduces four outputs: (i) S+L, (ii) S-L, (iii) S+jL,and (iv) S-jL. Each optical output pair, (i)–(ii) and (iii)–(iv), is collected by a pair of matched photodiodes whose photocurrents are subtracted to produce output currentsproportional to |S+L|2 - |S-L|2 = 4 · Re{SL*} and|S+jL|2 - |S-jL|2 = 4 · Im{SL*}; together, theseconstruct the complex value SL*.

INVITED PAPER

FIELDABLE DIGITAL COHERENT INTERFEROMETRIC COMMUNICATION AND SENSING APPLICATION DOMAINS

Abstract—A layered architecture unifying optical coherent communications and interferometric sensing via adigitally stabilized quadrature modulator and a homodyne receiver, augmented by digital noise reduction andchannel compensation algorithm stacks, forms a new paradigm for coherent fieldable applications.

Key Words—coherent communications, integrated optics devices, laser sensors


Isaac Shpantzer, PhDCeLight, [email protected]

____________________________

1 CeLight specializes in the research, development, andsales of advanced opto-electronic subsystems, algorithms,and architectures used in standoff detection of explosives,chemicals, and trace gases; secure and robust opticalcommunications; coherent LADAR and vibrometry; targetidentification, discrimination, and tracking; opto-electronicwarfare and signal intelligence; and medical diagnostics.

S=Ase j(ωt+φ )H

L=Ale jωtH

Signal

Local Oscillator

AmplitudeFrequency

PhasePolarization

Optical Hybrid S+L

S+jL

S-jL

S-L

BalancedDetectorsand TIAs

R{SL*}= AlAscos(φ )

I{SL*}= AlAssin(φ )

I(t)

Q(t)

Signal Boostedby LocalOscillator

Analog-to-DigitalConversion

Digital Processing

Advanced Algorithms

.

.

.

Phase{

Figure 1. Digital Coherent Integrated Homodyne Receiver for Linear Transformation of Encoded Optical Signal to Baseband (Top view shows fabricated device; tested device is shown at bottom right. For simplicity, only one polarization is shown.)

Following this linear transformation, the signalsare electrically filtered, sampled, and thenprocessed digitally as described in Reference [1].The key advantages of this coherent detectionscheme are:

• The received signal is boosted by the localoscillator to obtain the highest shot-noiselimited receiver sensitivity.

• Using an agile local oscillator makes thereceiver inherently frequency-selective.

• Linear down-conversion to electricalbaseband enables the use of adaptive DSP-based noise reduction algorithms.

The linear transformation allows the order ofcompensation to be changed, thus enabling theuse of back-end digital adaptive algorithms tocompensate for a variety of noise sources. Using these algorithms eliminates the need for traditional complex front-end opticaltechniques such as optical phase locking and polarization compensation. Furthermore,adaptive stabilization of the interferometriccomponents to maintain their operational pointsis solved (see later discussion in this paper andReference [2]).

The key benefit of the linear transformationdescribed above results from the ability tooperate on the field vector, thus enabling digitaladaptive compensation for multiplicative phasenoise induced by physical phenomena such asplatform vibrations, Doppler shift, polarizationrotation and birefringence, high-speed-air-turbulence-induced fading and scintillation thatcannot be compensated for by adaptive optics,and electronic beam-steering. The advantages ofthis approach are discussed later in this paperand in Reference [1].

LAYERED ARCHITECTURE FOR DIGITALCOHERENT COMMUNICATIONS AND SENSING

Aunified layered architecture for fieldabledigital coherent interferometric communi-

cations and sensing is depicted in Figure 2. Thelayered architecture is based on two integratedoptical components that enable the embodiment


Figure 2. Unified Layered Architecture for Fieldable Digital Coherent Interferometric Communications and Sensing


ASIC application-specific integrated circuit

BER bit error rate

D/As digital to analog (converters)

DSP digital signal processor

DWDM dense wavelength division multiplexing

FPGA field-programmable gate array

IED improvised explosive device

LADAR laser radar

M-PSK M-ary PSK

M-QAM M-ary quadrature amplitude modulation

OCDMA orthogonal code division multiple access

PSK phase shift keying

QM quadrature modulator

QPSK quadrature PSK

RF radio frequency

Rx receiver

SNR signal-to-noise ratio

TIA trans-impedance amplifier

The key benefit of the linear

transformationdescribed aboveresults from the

ability to operate on the field vector,

thus enabling digital adaptive

compensation formultiplicative phase

noise induced by physical

phenomena such asplatform vibrations,

Doppler shift,and more.

Systems

Subsystems

IntegratedComponents

Synthesizer Analyzer

Adaptive Noise Reduction Noise Reduction Algorithms

Stabilization Algorithms Stabilization AlgorithmsQuadrature Modulator Homodyne Receiver

Vibration, Channel Estimation,Doppler, Scintillation,

Polarization Scrambling

Vibration, Channel Estimation,Doppler, Scintillation,Polarization Tracking,

Beam Steering

of a generalized transponder consisting ofsynthesizer and analyzer constructs. Eachconstruct consists of three layers:

• An optical layer composed of an integratedquadrature modulator (QM) and homodynereceiver that performs the lineartransformation of coherent optical signalsto/from the electrical baseband

• A stabilization layer that maintains theoptical components at an optimizedoperating point

• Adaptive DSP-based noise cancellation tocompensate for multiplicative phase noiseresulting from platform vibration, Dopplershift, polarization rotation, and fading andscintillation, as detailed later in this paperand in Reference [1]

For free-space optical communications, thesynthesizer provides agile synthesis of key-basedmultidimensional hopping in time, along with frequency, polarization state, coherentmodulation scheme (e.g., M-ary phase shiftkeying [M-PSK], M-ary quadrature amplitudemodulation [M-QAM]), and symbol rate.Together, these factors produce an optimalcombination of security against jamming andeavesdropping, spectral efficiency, and shieldingfrom atmospheric conditions in the tacticalenvironment. In the latter application, theanalyzer embodies coherent detection of ageneralized key-based multidimensionallyhopped coherent optical signal via a generalizedhomodyne polarization diversity receiver withDSP-based adaptive algorithms that digitallyextract the information content from both channelnoise and key-based multidimensional optical

scrambling without using optical frequency andpolarization tracking and unwinding.Furthermore, for improved security andflexibility, this approach does not require a fixeddense wavelength division multiplexing(DWDM)-like channel structure; instead, it takesadvantage of the contiguous gridless selection ofany optical carrier frequency based onatmospheric conditions and tactical operationalneeds (as further described in Reference [1]).

High-spectral-efficiency-fiber-based coherentcommunications using CeLight’s components aredescribed in References [3] and [4]; an orthogonalcode division multiple access (OCDMA)approach is presented in Reference [5]. Co-locating the synthesizer and analyzer andsharing a common optical local oscillator enablesthe design of interferometric sensing applicationssuch as coherent laser radar (LADAR) orvibrometry using a common layered architecturewith the unique ability to switch applications viasoftware control.

STABILIZATION OF INTEGRATEDINTERFEROMETRIC OPTICAL COMPONENTS

CeLight designed, fabricated, and tested bothinterferometric components (QM and

homodyne receiver) with closed loopstabilization of their operating point foruninterrupted up/down linear conversion of baseband electrical signals of up to 12.5 Gsymbols/second. A stabilization algorithmfor the QM is described in Reference [2] and shown in Figure 3. Both simulation and experimental results for generating 12.5 Gsymbol/second quadrature PSK (QPSK)


0

30

6090

120

150

180

210

240270

300

330

0

30

6090

120

150

180

210

240270

300

330

0.5 0.2

0.4

0.6

0.8

1.0

1.5

(a) (b)

0 10 20 30 40 50

0.5

0.4

0.3

0.2

0.1

0

–0.1

–0.2

Bias

es an

d Ph

ase E

rror (

π)

Iteration Number

Bias 1 - πBias 2 - πPhase - π/2

2.0 1.0

(c)

Figure 3. (a) Constellation Plots of QM Optical Output at Startup; (b) Constellation Plots After 50 Iterations of the Control Loop;(c) Deviations of the Two Biases and the Phase from Their Optimal Points (ππ and ππ/2) Versus Iteration Number

A unified layeredarchitecture forfieldable digital

coherentinterferometric

communicationsand sensing is the

key for a variety of application

domains.


signals are described in Reference [2]. Thestabilization algorithm shows a 1-dB sensitivitypenalty compared with manually adjusting theQM by minimizing the bit error rate (BER).CeLight has developed and tested similarstabilization algorithms for the homodynereceiver; these will be reported in another venue. Further improvement of the stabilizationalgorithms is planned to accommodateinterferometric sensing.

ADAPTIVE NOISE REDUCTION ALGORITHMS

Reference [1] discusses digital adaptive noisereduction algorithms that compensate for

various physical phenomena. Figure 4(a)schematically shows the architectural approach asapplied to channel compensation of air turbulencefor free-space coherent QPSK communications.Figure 4(b) shows simulated results of a free-space QPSK optical link operating over ahigh-speed turbulent channel before and afteradaptive noise reduction, as well as the timeconvergence track.

INTERFEROMETRIC SENSING APPLICATIONS

Co-locating the synthesizer and analyzer withan optical local oscillator derived from the

transmitting laser turns the transponderconfiguration into an interferometric sensingsystem that enables a variety of applicationdomains, all of which use a unified layered

architecture with the unique ability to adapt,reconfigure, and switch coherent applications viasoftware control. The key application domains for digital coherent interferometric sensinginvestigated by CeLight include:

• Standoff Explosive Detection: The standoffdetection of improvised explosive devices(IEDs) by exciting their emitted trace gasesand interferometrically measuring the minuterefractive index phase changes of air inducedby the photothermal effect.

• Standoff Chemical Warfare Detection andChemical Plant Monitoring: Similar tostandoff explosive detection but using atunable laser to excite the trace gas molecules.

• Coherent Optical Fence: The pinpointing ofthe incident location of an intrusion bysensing the minute phase changes along afiber (laid underground to protect theperimeter of a valuable installation) thatresult from vibrations, magnetic field,electrical field, temperature, etc. This is theequivalent of a multisensor distributed fenceusing common communications fiber.

• 3-D Coherent Vibrometry: The ability toidentify typical Eigen-vibrations bymeasuring the minute phase changesresulting from target vibrations in threedimensions. This enables the discriminationbetween dummy and real nuclear warheadsin ballistic space or the identification ofvehicles (friend or foe) on the ground.

Figure 4. (a) Embodiment of Digital Adaptive Algorithms; (b) Performance of Channel Equalization Algorithm over Turbulent Atmospheric Channel

Co-locating thesynthesizer andanalyzer with an

optical localoscillator derived

from thetransmitting laser

turns thetransponder

configuration intoan interferometric

sensing system thatenables a variety of

applicationdomains.

2.5

OpticalInput Homodyne Receiver

High-SpeedASIC(XWH)

D/As

Optical Sampling

Rx Stabilization

X=IV+jQV+IHjQH

X=IV+jQV+IHjQH X=Iv+jQv+IHjQH

Filter Coefficients

W H

DSP/FPGA(Noise Reduction @

AtmosphericTurbulence Rate)

Rx Stabilization0

30

6090

120

150

180

210

240270

300

330

0

30

6090

120

150

180

210

240270

300

330

0

30

6090

120

150

180

210

240270

300

330

1.0

0.8

0.6

0.4

0.20 20 40 60

Number of Iterations

Mea

n Squ

ared

Erro

r, |e

(n)|

2

Convergence

SNR = 30

Equalized Symbols

Transmitted Samples Received Samples

~

5.0

2.5

5.0

2.5

5.0


CONCLUSIONS

Aunified architecture for fieldable opticalcoherent interferometric sensing was

developed and tested in a variety of applicationdomains. Adaptive DSP algorithms were used tostabilize the interferometric components andcompensate noise sources of the optical path,vibrations, movements and more, thus enablingpractical deployment.

ACKNOWLEDGMENTS

The author would like to thank the CeLight team—P. Cho, A. Kaplan, Y. Achiam,

S. Kazi, A. Greenblatt, G. Harston, J. Khurgin, andA. Salamon—for their valuable contributions tothe paper.

REFERENCES

[1] A. Salamon, G. Levy-Yurista, M. Tseytlin, P.S. Cho, and I. Shpantzer, “Secure OpticalCommunications Utilizing PSK Modulation,Polarization Multiplexing and Coherent DigitalHomodyne Detection with Wavelength and Polarization Agility,” MILCOM 2003, Session U026, Boston, MA, October 13-16, 2003.

[2] P.S. Cho, J.B. Khurgin, and I. Shpantzer, “Closed-loop Control of LiNbO3 QuadratureModulator for Coherent Communications,” COTA Conference Proceedings, Whistler, Canada,June 2006.

[3] P.S. Cho, G. Harston, C.J. Kerr, A. Greenblatt, A.S. Kaplan, Y. Achiam, and I. Shpantzer,“Coherent Homodyne Detection of BPSK SignalsUsing Time-Gated Amplification and LiNbO3Optical 90o Hybrid,” IEEE Photonics TechnologyLetters, Vol. 16, No. 7, July 2004, pp. 1727–1729.

[4] P.S. Cho, G. Harston, C.J. Kerr, A. Greenblatt, A.S. Kaplan, Y. Achiam, G. Levy-Yurista, M. Margalit, Y. Gross, and J.B. Khurgin,“Investigation of 2-bit/s/Hz 40-Gb/s DWDMTransmission over 4×100-km SMF-28 Fiber UsingRZ-DQPSK and Polarization Multiplexing,” IEEE Photonics Technology Letters, Vol. 16, No. 2,February 2004, pp. 656–658.

[5] J.B. Khurgin, A.B. Cooper III, P.S. Cho, and I. Shpantzer, “Painless Fully Orthogonal Coherent OCDM,” COTA Conference Proceedings,Whistler, Canada, June 2006.

BIOGRAPHYIsaac Shpantzer, founder andCTO of CeLight, is a technologyvisionary and accomplishedinventor with more than 37 years of advancedengineering achievements. Prior to his involvement withCeLight, Dr. Shpantzer wasfounder and president ofNextNet, where he developed a

carrier-class wireless data communications systemdeployed by service providers around the world, andexecutive vice president of engineering at Racotek,where he developed a mobile communications systemwith tens of thousands of stations deployed. Previously,he developed technologies and products related to fiberoptics, electronic warfare, vector signal processing,command, and control, and managed the SystemsDivision of Israel’s Nuclear Research Center.

Dr. Shpantzer has BS and MS degrees in ElectricalEngineering from Technion – Israel Institute ofTechnology, Haifa, Israel, and a PhD from theUniversity of Manitoba, Canada.


INTRODUCTION

Solar energy has proven to be a reliable and economical method of powering

telecommunication systems in places whereconventional electricity is unavailable orimpractical. It provides an excellent source ofclean, reliable power to keep batteries charged.This paper contains a brief description ofphotovoltaic (PV) cells and provides examples ofnew technologies for which solar power is—if notthe only option—an excellent choice of energy.

As the world's telecommunication networks areextended and upgraded, rural communicationservices come into greater focus. Because siteaccess is often difficult and connection to a central electricity grid is seldom possible, a stand-alone power system is required in rural areas.Such a system must be cost effective, simple tomaintain, and reliable.As the importance ofwireless communicationtechnologies grows, newdemands are placed on solar power. Solar power is also playing an important role inpowering the world’sconverging telecommu-nication markets.

The relative ease ofinstallation and indivi-dual ownership of solarenergy systems havemade them very popularfor receiving cellular,

radio, and television (TV) services in rural areas.This growing popularity is also attributable to thefact that people living in or traveling to manyremotely located rural areas simply have no otherchoice in service.

ELECTRICITY GENERATION THROUGH SOLAR CELLS

PV or solar cells are semiconductor devicesthat convert sunlight into direct current

electricity, bypassing thermodynamic cycles andmechanical generators. When light photons ofsufficient energy strike a solar cell, they knockelectrons free in the silicon crystal structure,forcing them through an external circuit and thenreturning them to the other side of the solar cell tostart the process all over again. See Figure 1.

SOLAR ENERGY IN TELECOMMUNICATIONS

Abstract—Solar energy systems are increasingly used for a variety of applications, among them newtelecommunication converging technologies, the focus of this paper. Several examples of the use of solar energysystems for different telecommunication scenarios are discussed. Because solar energy systems are robust andcost effective, the demand for solar power in telecommunications will continue to grow.

Key Words—insolation, irradiance, mesh network, Niue, photovoltaic, solar cell, solar energy, solar Wi-Fi,solar WiMAX, telecommunication


Glenn A. [email protected]

Mansour [email protected]

Hole

Electron

Cover Glass

n-Type Semiconductorp-Type Semiconductor Back Contact

Transparent Adhesive

Antireflection Coating

Front Contact

Sunlight

Current

Figure 1. Typical Solar Cell Generator

The voltage output from a single crystalline solarcell is about 0.5 volt with an amperage output thatis directly proportional to the cell's surface area(approximately 7 amperes for a 6-inch-squaremulticrystalline solar cell). Typically, 30 to 36cells are wired in series (+ to –) in each solarmodule. This produces a solar module with a 12-volt nominal output (~17 volts at peak power)that can then be wired in series and/or parallelwith other solar modules to form a complete solar array.

Figure 2 shows the components needed for asolar energy system.

Solar is universal and will work virtuallyanywhere; however, some locations are obviouslymore suitable than others. Irradiance is a measureof the sun's power available at the earth’s surface,with power peaking at about 1,000 watts persquare meter (W/m2). With typical crystallinesolar cell efficiencies around 14 to 16 percent,about 140 to 160 watts can be generated persquare meter of solar cells placed in full sun.



AH ampere hourAP access pointCAPEX capital expendituresCPE customer premises equipmentGPS global positioning systemHAP high altitude platformIEEE Institute of Electrical and

Electronics EngineersIP Internet ProtocolLOLP loss-of-load probabilityNREL National Renewable Energy

LaboratoryOPEX operating expendituresPC personal computerPDA personal digital assistantPV photovoltaicRF radio frequency

140 to 160 wattscan be generatedper square meter

of solar cells placed in full sun,

with typicalcrystalline solar cell efficiencies

around 14 to 16 percent.

SSID service set identifierTV televisionUHF ultra high frequencyVHF very high frequencyVoIP voice over IPWi-Fi® wireless fidelity

(Although synonymous with the IEEE 802.11 standards suite and standardized by IEEE, Wi-Fi is a certification mark promoted by the Wi-Fi Alliance.)

WiMAX™ Worldwide Interoperability for Microwave Access (Although synonymous with the IEEE 802.16 standards suiteand standardized by IEEE, WiMAX is a certification mark promoted by the WiMAX Forum.)

Figure 2. Solar Energy System Components

BatteryInverter

AC Loads DC Loads

ChargeController

Solar Irradiance

Module


Insolation, which is a measure of the availableenergy from the sun, is expressed in terms of “full sun hours” (i.e., 4 full sun hours = 4 hours ofsunlight at an irradiance level of 1,000 W/m2).Obviously, different parts of the world receivemore sunlight than others, so these areas havemore full sun hours per day. The solar insolationzone map provides a general idea of the full sunhours per day during the year for locations in thecontinental US. See Figure 3.

WIRELESS FIDELITY

FunctionalityIn popular usage, the term Wi-Fi® has beenadopted to refer to wireless fidelity technology.This technology is supported by the Institute ofElectrical and Electronics Engineers (IEEE) 802.11family of standards. Products tested andapproved by the Wi-Fi Alliance® as Wi-FiCERTIFIED™ are interoperable with each other,even if from different manufacturers, and areallowed to display the Wi-Fi certification mark. A user with a Wi-Fi CERTIFIED product can use any brand of access point (AP) with any otherbrand of client hardware that is also certified.Typically, in fact, any wireless fidelity 2.4 GHz(for 802.11b or 802.11g) and 5 GHz (for 802.11a)product using the same radio frequency (RF) can work with any other, even if not Wi-Fi CERTIFIED.

A person with a Wi-Fi-enabled device such as acomputer, cell phone, or personal digital assistant(PDA) can connect to the Internet when in the

range of an AP. An AP broadcasts its service setidentifier (SSID) via packets called beacons. Thebeacons are transmitted at 1 Mbit/s every 100 ms.

Solar-Powered Wi-Fi

Niue’s Solar-Powered Wi-Fi SolutionNiue, a speck of land in the middle of the SouthPacific Ocean, is the world's tiniest island nation.The people of Hakupu, Niue's second largestvillage located about 7 miles from the capitalAlofi (see map in Figure 4), were dissatisfied withthe dial-up Internet connections availablethrough their existing telephone system. Theancient phone system at best could handle

Figure 3. Average Daily Solar Radiation per Month (Annual) [1]

Two-Axis Tracking Flat Plate

10 to 148 to 107 to 86 to 75 to 64 to 53 to 42 to 30 to 2none

kWh/m2/day

Niue's most-isolated places

can have solar-powered Wi-Fi hotspots

that allow net surfing with

much greater speedand reliability, as in Hakupu.

Figure 4. Map of Niue Island [2]

169° 50’ N 169° 45’ N

19° 00’ N

19° 05’ N

00

3 Miles3 Kilometers

Mutalau

Lakepa

Liku

HakupuAvateleVaiea

Makefu

Toi

TuapaNamukulu

Hikutavake

Tarnakautonga

AlofiNiue

(New Zealand)

PacificOcean

169° 55’ N


4,800 baud over dial-up service. By 2005,Hakupu's Internet users were looking to local Wi-Fi as a solution, though they were 7 milesdistant from a satellite earth station, the nearest AP.

Fortunately for the people of Hakupu, anabandoned tower nearby was identified assuitable for use in bringing Wi-Fi coverage fromthe earth station 7 miles away. The 120-foottower, with no electricity at the site, had seemedto hold little prospect for useful service. Instead,the circumstances became the perfect equation

for the tower installation to gain new life as a solar-powered Wi-Fi station. Hakupu’s newfound ability to surf the net with muchgreater speed and reliability soon led to requestsfrom other outlying villages in Niue for similarservice. Solar-powered units like Hakupu’s could be the answer.

Figure 5 shows the basic wiring diagram for the solar-powered Wi-Fi repeater station and AP in Niue [2].

St. Louis Park, Minnesota—First Solar-Powered Wi-Fi Community in the USSt. Louis Park, a Minneapolis suburb of about44,000, recently became the first US community toagree to city-wide deployment of solar-poweredWi-Fi service. After a phased-in installation, theentire network is slated to be on line by July 2007 [3]. St. Louis Park’s service will bepowered by about 400 solar panels suspended 20 to 30 feet in the air on public rights-of-way,such as roadsides. Each panel is 805 squareinches—about the size of a stop sign, butrectangular and black. See Figure 6.

Solar power mitigates reliance on the electricalgrid. Especially during storm-related emergencies,St. Louis Park's solar-powered wireless APs, inconjunction with battery-powered backup at allpoints along the network, will provide real valuefor public safety and services applications. Thanksto solar power, those responsible for public safety

Figure 6. Solar-Powered Wi-Fi AP on a Utility Pole [3]

Solar Module

Wi-Fi

Utility Pole

Figure 5. Basic Wiring Diagram for the Solar-powered Wi-Fi Repeater Station and AP in Niue [2]

Charger12 Volt

Gell Cell Sealed150 AH 12 Volt

Backup PowerFrom Mains(Optional)240 VAC

Regulator12 Volt

Inverter

Tower

Wi-FiEquipment

Solar Panel60 Watt

Solar Panel60 Watt

With solar-poweredwireless APs,

those responsiblefor public safety

and services can continue to usethe network during

electric gridoutages.


and services can continue to use the networkduring electric grid outages.

As shown in Figure 7, the network architecturewill support either router- or bridge-basedcustomer premises equipment (CPE).

The design of the entire solar power supply is based on the historic climatic data forMinneapolis, as provided by the NationalRenewable Energy Laboratory (NREL).

The wireless APs will be housed in weatherproofenclosures that incorporate needed components,including a rechargeable battery with a 5.5-daypower supply (for use in the event of solar panel destruction).

The ultimate indication of the system’s reliabilityis its loss-of-load probability (LOLP). The actualLOLP for December (the worst month) is 0.05 percent, which equals about one Decemberday every 64 years [4].

Solar Wi-Fi Plus Satellite Backbone for Rural Areas Solar Wi-Fi plus satellite on a mesh network canbe a good choice for wireless communication inrural areas. In a mesh network, each user cancommunicate with the AP and also directly withother users. If the AP is out of range, intermediateusers can relay messages to the AP.

A benefit of this method is that it extends therange of the network without resorting to higherpowers or external antennas. A limitation is that aline of sight is still needed between the usersmaking up the chain. If an intermediate terminal

goes down, all other subscribers connectingthrough that terminal are cut off. In practice, thenumber of hops from any subscriber to the APshould be kept to three.

Mesh products are not commonly available atpresent, but they hold the promise of “free” range extension.

The number of subscribers on a single channelshould stay within the range of 20 to 30 to sustainusable throughput. Assuming a usable range of2.5 km from the AP, this number equates to asubscriber density of 1 to 1.5 users per km2.

OTHER SOLAR-RELATED TELECOMS DEVELOPMENTS

Solar Power and WiMAX™

WiMAX DefinedIn popular usage, the term WiMAX™ has beenadopted to refer to the technology behindWorldwide Interoperability for Microwave Access.This technology is supported by the IEEE 802.16family of standards. Projects tested and approvedby the WiMAX Forum™ as WiMAX ForumCertified™ are interoperable with each other, evenif from different manufacturers, and are allowed todisplay the WiMAX certification mark.

Solar-Powered Pre-WiMAX System—Lake Tahoe, NevadaAirTegrity Wireless™ has donated a pre-WiMAXsystem (WiMAX-in-a-Box™) to deliver wireless

Figure 7. Logical Network Diagram [3]

WirelessNetwork

Admin IPCPE Bridge

PC Has Network IP

PC MustAuthenticate

Admin IP

CPEAuthenticator

802.11x

CPE HasNetwork IP

PCNon-Routing

IP

Provided by Designated

St. Louis ParkManagement

Partner

CPE Router

L2 Authenticator

A solar-poweredpre-WiMAX

base stationdelivers wireless

broadband to Sand Harbor in theLake Tahoe-Nevada

State Park.


broadband to Sand Harbor in the Lake Tahoe-Nevada State Park. This park is home to the LakeTahoe Shakespeare Festival every summer, andwireless Internet access will now be available tothe park’s employees, volunteers, and visitors.The solar-powered pre-WiMAX base station isinstalled at an elevation of 8,000 feet, making it the highest elevation pre-WiMAX deploymentin the nation and subjecting it to harshenvironmental conditions. See Figure 8.

Until certification is completed, the AirTegrityWireless system should be considered a pre-WiMAX technology because it is based on thecurrent specifications for WiMAX.

Stratellite™ AirshipStratellite is a massive solar-powered airship thatwill be used to transmit wireless communicationsat 65,000 feet above the Earth. This specializedairship (see Figure 9) is being built by SanswireNetworks. Under current development programs,delivery of a demonstration airship is targeted forthe 2008–2011 timeframe [6].

Based on the frequency band used, Stratellite willcover a 500 km radius footprint for very highfrequency (VHF). If ultra high frequency (UHF) isavailable, a 250 km footprint is possible. At 3.5 GHz, the radius would be 75 km. A thin-filmPV array will be located on the top of the airship.Batteries will be used through the night cycle. Theairship will float using the buoyancy of the liftinggas, assisted by the propulsion units. It will be

global positioning system (GPS) guided andground controlled. The airship will remain aloftfor up to 18 months, and bringing it down should not be difficult [6].

The Stratellite network will obtain its backhaulover fiber connections. Point-to-point links willbring the bandwidth up to the airship. Routersand switches within the airship will assign thebandwidth to the different transponders. Point-to-multipoint radios will distribute the bandwidth to the CPE units on the ground. Optional groundrepeaters will act as fill-in units, where necessary.If multiple airships are used in a region, inter-airship communication will provide additionalredundancy. See Figure 10.

The high altitude platform (HAP) conceptembodied by Stratellite is not new. While the idea has been around for many years, Sanswireindicates that recent airship development has been assisted by advances in lightweight materialsand by new power possibilities stemming fromadvances in solar technology, energy storage,energy management, and propulsion.

THE BUSINESS CASE FOR SOLAR

Studies have shown that solar energy isbecoming an increasingly viable alternative

[7, 8] where there is no access to the electricitygrid, providing the only realistic solution in somesituations, as discussed earlier in this paper. Evenwhen access to the electricity grid is not an issue,a solar energy network can still make sense. As anexample, the discussion returns to St. Louis Park,described earlier in this paper. The business casefor this community [9], the first solar-poweredWi-Fi locality in the US, is summarized below [4].As the numbers show, a solar energy network isstill a very viable solution even with access toconventional grid electricity because solar- andbattery-powered radios eliminate the energyportion of the pole attachment fees.

Figure 9. Stratellite Airship [6]

Airship developmenthas been assisted

by advances inlightweightmaterials

and by new power

possibilitiesstemming from

advances in solar technology,energy storage,

energymanagement,

and propulsion.

Figure 8. Pre-WiMAX Station at Sand Harbor [5]


Capital Expenditures (CAPEX)

Wireless network equipment/installation $1,144,321

Fiber network/backhaul (20 years of service) 483,300

CPE (1,000 units) 86,000

Additional implementation costs include:

Storage, facility upgrades, project management, and miscellaneous expenses 150,000

Contingency and 50 percent of pilot costs 225,000

Equipment spares 106,000

Core network (routers, switches, and authentication) 200,000

Sales tax 190,000

CAPEX total $2,584,621

Operating Expenditures (OPEX)

Wi-Fi (5 years of operation) $2,845,403

Given the above assumptions, the break-even cash flow is at a subscriber level of 32 percent—6,100—of all eligible households at a subscriptionfee of $14 per month. According to Figure 11, a 32 percent subscriber penetration rate is arealistic goal.

CONCLUSIONS

Solar power has proven a reliable powersupply for telecommunications. The techno-

logy has been shown to be robust and costeffective where conventional grid electricity is notavailable. With networks expanding into ruralareas using wireless technology, the demand forsolar power will continue to grow. The fallingcost of solar technology and the increasedsophistication of control and supervisionequipment will ensure that solar power remainscommercially and technically competitive.

16% 23%

39%

53%

73%

84%

20%

63%

51%

27%

33%

33%

17%

11%6%

10%14%

26%

Likely

Up to $20 perMonth (N=277)

$21 to $25 perMonth (N=278)




$41 or more perMonth (N=277)

Neutral

Not Likely

Figure 11. Subscriber Willingness to Switch to St. Louis Park Wireless Service [10]

Figure 10. Typical Stratellite Network [6]

IP-BasedCellular Phone

GPS Satellite

Sanswire HAPStratellite

Stratellite Force MultiplierCommunications

CommunicationsSatellite

IP/Hi-DefTelevision

Hi-Speed IPBroadband Service

PDA“All-In-One”

Device

Digital EnhancedCordless

Telecommunications

VoIPTerrestrial

InfrastructureAugmentation

Network Operations CenterCommand and Control

Uplink/Downlink


TRADEMARKS

AirTegrity Wireless and WiMAX-in-a-Box aretrademarks of AirTegrity Wireless, Inc.

Stratellite is a trademark of Sanswire Networks.

Wi-Fi and Wi-Fi Alliance are registeredtrademarks and Wi-Fi CERTIFIED is atrademark of the Wi-Fi Alliance.

WiMAX, WiMAX Forum, and WiMAX ForumCertified are trademarks of the WiMAX Forum.

REFERENCES

[1] “US Insolation Maps: Fixed Azimuth, Single andDual-Axis Tracked Flat Plates,” Watson™ SolarTrackers (http://www.wattsun.com/resources/insolation_maps/map_index.html).

[2] R. St. Clair, “Solarfi: Niue’s WiFi Nation Goes Green,” Internet Users Society – Niue white paper, Number 4, July 2005(http://www.nunames.nu/ about/SolarFi.PDF).

[3] Wireless Internet Service Provider (for Residents),City of St. Louis Park (http://www.stlouispark.org/residents/wireless.htm).

[4] St. Louis Park Council Meeting, Item: 110606 – 8d – Citywide Wireless Project, p. 4(http://www.stlouispark.org/pdf/110606_8d_Citywide_Wireless_Project.pdf).

[5] “AirTegrity Wireless Donates Pre-WiMAXSolution to Sand Harbor State Park to ProvidePark and Lake Tahoe Shakespeare Festival withFree Wireless Internet Access,” AirTegrityWireless press release, August 26, 2006(http://www.airtegrity.com/media_center/press_releases).

[6] Sanswire Networks(http://www.sanswire.com/).

[7] Global Environment & Technology Foundation(http://www.getf.org).

[8] Australian Greenhouse Office(http://[email protected]).

[9] City Council Study Session, Report Item 071706 – 2 – Wireless Pilot Update(http://www.stlouispark.org/pdf/July_17_2006.pdf).

[10] City Council Study Session, Report Item 071006 – 6 – Wireless Pilot Update(http://www.stlouispark.org/pdf/July_10_Wireless_Pilot_Update.pdf).

BIOGRAPHIESGlenn Torshizi joined BechtelTelecommunications in 2001and is currently working assenior RF lead design engineeron the Bechtel MSV WiMAXATC network preliminarydesign for the Baltimore,Washington, New York, Boston,and Miami markets. Before this, he worked on the Modeo

DVB-H project as a senior RF design engineer. He has finished the preliminary SFN DVB-H design for Boston; New York expansion; Philadelphia;Baltimore; Washington, DC; and Pittsburgh. Before this,he was a staff scientist/engineer at the BechtelTelecommunications Training, Demonstration, andResearch (TDR) Laboratory in Frederick, Maryland. For 1 year, Glenn was involved with Bechtel’s VirtualSurvey Tool (VST) and received the award for mostinnovative exhibit at Bechtel’s sixth FrederickTechnology Fair. Glenn also finished an evaluation test on a 2.4 GHz Wi-FiTM phased array antenna and published the results in the January 2006 issue ofthe Bechtel Telecommunications Technical Journal. Beforerelocating to Frederick, he spent more than 3 years as anRF design engineer on the Bechtel AWS GSM, GPRS,and UMTS Program in Philadelphia, Pennsylvania, andas the market RF lead in Harrisburg, Pennsylvania, and Hackensack, New Jersey.

Before joining Bechtel, Glenn helped plan, optimize,and integrate the Triton PCS TDMA system in Norfolk, Virginia, and the Cricket CommunicationsCDMA system in Pittsburgh, Pennsylvania. As atechnical expert witness on numerous planning andzoning boards, he was very successful in obtaining final site approvals.

Glenn has a BS in Physics from SouthwesternOklahoma State University and an MS in Physics from the University of Tennessee, Knoxville. He has done research in relativistic heavy ion physics atOak Ridge National Laboratory and BrookhavenNational Laboratory.

Mansour Niknam, currently a manager in the NetworkPlanning department of Bechtel Telecommunications, isresponsible for staffing, qualityassurance and control, planningtools, and training. Previously,he was a senior project controlsengineer on the AT&T Libertywireless projects. Mansour

joined Bechtel in 1997 as a senior project controlsengineer in the Power organization. In this capacity, hemaintained master budgets and schedules andsummary schedules for combined cycle steam powerplant construction projects; analyzed, forecasted, andreported productivity and performance trends; andmonitored quantity deviations, among other schedulingrelated duties.


For most of his career before joining Bechtel, Mansourwas involved as an engineer or manager in the designand construction of power generation and industrialwater treatment projects. He was employed over an 8-year period by several environmental engineeringcompanies in the Netherlands to manage alternativeenergy (waste, solar, wind) and oil refinery/petrochemical plant water treatment projects in morethan 10 countries. He helped develop the use of wasteand refuse-derived fuel to generate power for urbanand rural applications in northern Europe. He hasparticipated in international conferences and preparedtechnical proposals on alternative energy and watertreatment and has actively contributed his technicalassistance to international organizations providingwater treatment technology and alternative energysources to developing countries. In the Republic of Suriname, for example, his technical expertise assisted local farmers in constructing fruit dryers using solar energy.

Mansour holds a BE in Marine Engineering from theState University of New York, New York MaritimeAcademy, and an MBA from Frostburg State Universityin Maryland.


WIRELESS LOCAL LOOP

Overview

Defining WLLThe local loop is the physical connection betweenthe terminating equipment in a central office (CO)and the end users. It is often provisioned as atwisted pair of copper wires. Traditionallyprovisioned for plain old telephone service(POTS) voice only, it is now being used to carry arange of other technologies such as integratedservices digital network (ISDN) and digitalsubscriber line (DSL). The term local loop issometimes applied to any “last-mile” connectionto the customer, regardless of technology orintended purpose. This makes sense movingforward with the shift to data services andbroadband wireless access (BWA). Therefore, awireless local loop (WLL), in its simplestconstruct, uses a wireless radio signal to replaceall or part of the traditional copper infrastructurebetween a subscriber and the switch.

A key point to highlight is that a WLL is anapplication, and not necessarily a specifictechnology. The technology is only an enabler forthis type of application. Various WLL systemsand technologies exist. They include, but are not

limited to, cordless access systems, proprietaryfixed radio systems, and fixed cellular systems.As technology has evolved, so has WLLimplementation. Today, the most common WLLimplementation is via code division multipleaccess (CDMA) cellular technology. Thistechnology is examined further in this paper.

The mobile technologies—global system for mobile communication (GSM)/universal mobiletelecommunications system (UMTS), CDMA/cdma2000®, etc.—can be deployed either in theirtraditional mobile topologies or in a morelimiting fixed topology. Fixed wireless terminal(FWT) units differ from conventional mobileterminal units operating within cellular networksin that an FWT is limited to an almost permanentlocation, with virtually no roaming capabilities.Also, an FWT is usually physically less conduciveto mobility and more suitable to fixedcommunication devices. These devices could be adesk telephone; an interface for a traditionalanalog telephone; a connection to a fax machine,terminal, or data card to provide Internetconnectivity; or even a combination of these orsome other more specialized interfaces. Figure 1shows a few common FWT devices, including acommon mobile device. [1]

cdma2000® WIRELESS LOCAL LOOP EVOLUTION AND PERFORMANCE

Abstract—WLL, once only a simple wireless replacement for copper wires, has evolved with the ever-changingtechnology and market demands. Through the evolution of CDMA to cdma2000®, EV-DO, and eventually EV-DO Revision C, and with performance approaching or equal to current DSL networks, WLL is no longerlimited to only voice service in remote regions. Fixed and mobile broadband wireless access is the new “WLL.”

Key Words—1xEV-DO, 1xEV-DO Revision A, 1xEV-DO Revision B, all-IP-based architecture, broadbandwireless access, cdma2000 evolution, CDMA/cdma2000, FWT, multicarrier CDMA, VoIP over 1xEV-DO,wireless local loop, WLL


Nathan T. [email protected]

Figure 1. Common FWT Devices



1xEV-DO 1x evolution-data optimized1xEV-DV 1x evolution-data/voice1xRTT 1x radio transmission

technology3G third generation enhanced

digital mobile phone service at broadband speeds enabling both voice and nonvoice data transfer

3x multicarrier CDMA using three 1.25 MHz carriers

3xMC-DO 3x multichannel-data optimized4G fourth generation enhanced

digital mobile phone service boosting transfer rates to 20 Mbps

A-key authentication keyBSC base station controllerBTS base transceiver stationBWA broadband wireless accessCAVE cellular authentication and

voice encryptionCDG CDMA development groupCDMA code division multiple accesscdma2000® A family of standards,

developed through compre-hensive proposals from QUALCOMM, describing the use of CDMA technology to meet 3G requirements for wireless communication systems

CMEA cellular message encryption algorithm

CO central officeCPE customer premises equipmentDL downlinkDSL digital subscriber lineESN electronic serial numberFDD frequency division duplexFTP file transfer protocolFWT fixed wireless terminalGPRS general packet radio serviceGPS global positioning systemGSM global system for mobile

communicationHARQ hybrid automatic repeat

requestHSDPA high-speed downlink packet

accessHSPA high-speed packet accessiDEN integrated digital enhanced

network

IMT-2000 International Mobile Telecommunications-2000

IP Internet ProtocolISDN integrated services digital

networkITPC Iraq Telephone and Post

CompanyLMDS local multipoint distribution

serviceLS-OFDMA layered superposed OFDMALTE long-term evolutionMAC media access controlMIMO multiple-input, multiple-outputMIN mobile identification numberMMDS multichannel, multipoint

distribution serviceMSC mobile switching centerOFDM orthogonal frequency division

multiplexing OFDMA orthogonal frequency division

multiple accessPDSN packet data service networkPMP point-to-multipointPN pseudo-random noisePOTS plain old telephone servicePSTN public switched telephone

networkPTP point-to-pointPTT push-to-talkQAM quadrature amplitude

modulationQoS quality of serviceRF radio frequencySDMA spatial division multiple accessSIP session initiation protocolSSD shared secret dataTDM time division multiplexingUL uplinkUMTS universal mobile

telecommunications systemVoIP voice over IPWiMAX™ Worldwide Interoperability

for Microwave Access (Although synonymous with the IEEE 802.16 standards suiteand standardized by IEEE, WiMAX is a certification mark promoted by the WiMAX Forum™.)

WLL wireless local loop

WLL AdvantagesSeveral advantages of WLL systems make theman attractive alternative to traditional outsideplant deployments. These advantages can becharacterized as cost, deployment speed, designconstraints, and flexibility. [2]

Cost—The main cost advantage of using wirelesstechnologies to deploy the “last mile” tosubscribers accrues from not having to deploypotentially costly outside plant infrastructure. Awireless site is limited by its coverage area andcapacity. Within these limits, there is a fixed costper subscriber. With copper and fiber cables, onthe other hand, the cost depends on the numberof subscribers and their distances from the CO, asshown in Figure 2. This fact makes a WLL anaffordable alternative to wireline in lowteledensity areas. A further cost improvementcomes with economy of scale. The cost of wirelessdrops as technologies advance and the number ofusers/subscribers increases. Conversely, the costof digging trenches for fiber cables remainsrelatively constant.

Deployment Speed—A wireless network can beinstalled and commissioned relatively quicklyand easily. Adding new users is also very easy. Auser needs only to purchase a suitable end-userdevice (phone, terminal, etc.) and sign up for acorresponding service plan. It is not necessary tosend field personnel to turn the service on. Inother words, no truck rolls are necessary, andadds, moves, and changes can be addressed froma centralized location.

Design Constraints—Three key drivers must beconsidered when designing WLL networks:voice quality, coverage, and capacity. Since these

three key drivers are always competing amongthemselves, it may be necessary to first determinean acceptable voice quality level and to thenchoose a correspondingly suitable wirelesscommunication system that can provide highcapacity and large coverage. [3]

Although adhering to the standard wirelessdesign criteria (such as coverage requirements,terrain conditions, capacity, and interference),WLL networks offer certain advantages over theirmobile counterparts. First, WLL deployments donot need to be contiguous. In other words, it isnot necessary to worry about handoff boundaries.Second, because of limited or no roaming, WLLnetworks provide increased coverage andcapacity over similar mobile networks. WLLnetwork coverage also increases to some extentbased on the type of end-user device. In manycases, end-user terminals and telephones arecapable of using a higher transmitter power andhave better receive sensitivities than mobiles.

Of course, an important design constraint iswhether or not roaming capabilities are enabled.Roaming can be enabled to any degree, fromlimited service to full mobility, generallydepending on the prevailing regulatoryenvironment.

Flexibility—The implementation of WLLnetworks is extremely flexible, especially withregard to expansion and increased capacity. It iseasy to expand a network to support addedcoverage and capacity requirements.

The fact that the process of building a WLLsystem does not require precise knowledge of theuser’s location adds flexibility to system planning


Figure 2. Cost of WLL Versus Outside Plant [1]

The main cost advantage

of using wireless

technologies to deploy

the “last mile” to subscribersaccrues from

not having to deploypotentially costly

outside plantinfrastructure.

1,500

1,000

500

0

Distance

Cost

($)

WirelessWireline


and deployment [2]. Moreover, adds, moves, andchanges are readily addressed in a wirelessnetwork and WLL devices can be easily moved toaccommodate new changes.

Depending on topology (urban, suburban, rural),WLL networks can support both large and smallcoverage areas and more readily cover adverseterrain than can a fixed infrastructure.

WLL Disadvantages Several issues need to be considered whendeciding to roll out a WLL network:

• A subscriber can only access the networkwhen receiving a suitable signal level.Therefore, it is important to design WLLnetworks for the appropriate coverage areas.

• Unlike a copper infrastructure, a WLLnetwork has no per-user fixed capacity andinstead shares the total capacity among users. Therefore, more caution is needed in designing for the expected capacityrequirements.

• Though they are expected to be lower thanthe cost of digging and trenching for a copperinfrastructure, the costs associated withconstructing and installing a WLL networkcan include tower construction or lease,backhaul, and even spectrum costs.

• Spectrum itself is a critical issue whendeploying a WLL network. What spectrum isavailable? What are the licensing costs? Arethere interference issues? What subscriberdevices are available? Spectrum affects thesystem’s available technologies, coverage,and capacity.

• Less critical, but still an issue, is the use ofspecialized customer premises equipment(CPE). Every WLL network user requires acustomer device, such as an FWT, to accessthe network. Specific requirements dependon the carrier, technology, and frequencyused.

One thing to keep in mind is that by leveragingthe economies of scale, the bigger the networkand the greater the number of subscribers, thesmaller the impact of these disadvantages.

WLL Growth and EvolutionBecause of the advantages of WLL and the lack ofexisting infrastructure, most traditional WLLgrowth is occurring in emerging economieswhere half the world’s population lacks basictelephone service. Countries such as China, India,Brazil, Russia, and Indonesia look to WLL

technology as an efficient way to deploy POTS formillions of subscribers without the expense ofburying tons of copper wire [1].

For the rest of the world, including theunderdeveloped countries, WLL technology isevolving to provide more than just basic voiceand data services; it is merging with broadbandwireless access enabled by third and fourthgeneration (3G) and (4G) network technologiessuch as high-speed downlink packet access(HSDPA), cdma2000 1x evolution-data optimized(1xEV-DO), and Worldwide Interoperability forMicrowave Access (WiMAX™).

WLL Technologies and Why CDMAAlthough a primary focus of WLL networks hasbeen to provide basic voice services, an additionalchallenge for operators is meeting the increasingdemand for Internet access and broadband dataapplications. It is obvious that previouslyoptional Internet access and data services arebecoming mandatory. Therefore, apart from thebasic coverage requirement, WLL networks must meet at least two additional minimumrequirements: sufficient voice capacity and dataservice [4]. This section discusses how this isbeing accomplished.

WLL TechnologiesWLL systems have traditionally been separatedinto two main groups: those operating above 2.4 GHz and those operating below 2.4 GHz. Thetechnologies that use frequencies above 2.4 GHzare usually able to support data speeds matchingISDN, cable, and DSL services, due to theavailability of greater bandwidths. Multichannel,multipoint distribution service (MMDS) fallsunder this group of technologies. Technologiesoperating above 20 GHz include those designatedas local multipoint distribution service (LMDS).These solutions are exclusively geared towarddata services and rely on point-to-point (PTP) andpoint-to-multipoint (PMP) primarily line-of-sightdeployments. In most cases they can support data transmission rates of several megabits persecond. However, radio coverage, equipmentcost, and availability are key limitations of these services. [4]

Below 2.4 GHz, the common air-interfacetechnologies include mostly mobile servicetechnologies such as GSM, CDMA, cdma2000,and UMTS. These systems are deployedthroughout the usual frequency bands (450, 800, 900, 1700, 1800, 1900, and 2100 MHz). Themost common frequency bands used for WLL are450, 800, and 1900 MHz. The 450 MHz band,

WLL technology isevolving to provide

more than just basic voice anddata services;

it is merging withbroadband wirelessaccess enabled by3G and 4G network

technologies such as HSDPA,

1xEV-DO, andWiMAX.


where available, is a popular choice in poor andunderdeveloped areas due to its increased radiofrequency (RF) propagation characteristics, asillustrated in Figure 3.

Since WLL networks are typically developedusing mobile cellular technologies, what makesthem different from traditional mobile services?In most cases, the only differentiation is the wayin which the wireless technology is implemented.The WLL operation is an alternativeconfiguration of fully mobile technology thatcomplies with the regulatory requirements of thecountry in which it is deployed. Implementationsrange from fixed configurations with no mobility(except within a single cell) to fully mobileconfigurations. This is due largely to theregulatory environment where it is deployed.

The Emergence of CDMAThrough the 1990s, CDMA, and now cdma2000,has emerged as the most widely deployed WLLset of standards. Besides available frequencyband allotments, other factors, such asinfrastructure and handset availability, cost,spectral efficiency, data service capabilities, andevolution strategies, have played a role in thetechnology selection. To many operators, CDMAWLL best addresses the demands and has helpedlay the foundation for a smooth migration to newand enhanced services [6]. Most WLL networks in the world are based on cdma2000 at either 450 MHz or 800 MHz [4].

CDMA Security—As with any network,especially wireless, security concerns are a keyissue. It is often said that “CDMA technology isinherently secure.” Although there are alwayssecurity-related issues, CDMA technology hasseveral inherent features that make it superior toits predecessors. This is in part based on the factthat CDMA technology originated from militaryapplications and cryptography. [7]

The inherent security of the CDMA air interfacecomes from its noise-like signature resulting fromthe use of spread spectrum technology and ofcodes. This signature makes the signal of any oneuser difficult to distinguish and decode withoutprevious knowledge and synchronization of thecodes being used.

CDMA uses specific spreading sequences andpseudo-random codes for the forward andreverse links. The spreading techniques formunique code channels for individual users in bothdirections of the communication channel. This isaccomplished by using the long code (a 242-bitpseudo-random noise [PN] sequence), the shortcode (a 215-bit PN sequence), and Walsh codes.Together, these codes scramble, spread, andidentify the users and channels in both directionsof the CDMA air interface [7]. However, simplyknowing the codes is of no value to the would-be eavesdropper because, without precise synchronization, the signal retains itsnoise-like characteristics. This essential need for

Figure 3. Frequency Coverage Areas [5]

Frequency(MHz)

Cell Radius(km)

RelativeCell Count

Cell Area(km2)

450

850

950

1800

1900

2100

48.9

29.4

26.9

14.0

13.3

12.0

7,521

2,712

2,269

618

553

449

1.0

2.8

3.3

12.2

13.6

16.2

GSM-1800IMT-2000

GSM-900CDMA450Coverage

Area


synchronization is the reason CDMA basestations require global positioning system (GPS) timing.

CDMA also has a unique soft-handoff capabilitythat allows a mobile to connect simultaneously toas many as six radios in the network, each with itsown Walsh code. This means that someoneattempting to eavesdrop on a subscriber’s call hasto have several devices connected at exactly thesame time in an attempt to synchronize with theintended signal. In addition, CDMA employs afast power control, 800 times per second, tomaintain its radio link. This characteristic alonemakes it difficult for a third party to achieve a stable link for interception of a CDMA voice channel, even with full knowledge of theWalsh code. [8]

In cdma2000 1xEV-DO, the forward link uses ratecontrol instead of power control and timedivision multiplexing (TDM) instead of spreadcodes. However, it still has inherent security thatprotects the identity of users and makesinterception very difficult. In addition, the mediaaccess control (MAC) identification assigned tousers is encrypted. User packets are assignedvariable time slots, and the data rate is controlledby the access terminal based on radio conditions.Packets are divided into subpackets using hybridautomatic repeat request (HARQ) and earlytermination mechanisms. These attributes make it virtually impossible to identify the user or to correlate user packets. The 1xEV-DO standard specification supports a securityprotocol layer that is ready to implement futuresecurity protocols. [7]

CDMA addresses encryption and authenticationby using a 64-bit authentication key (A-key) andthe electronic serial number (ESN) of the mobile.Together, these are used to generate subkeys thatprovide voice privacy and message encryption.

CDMA uses the standardized cellular authenti-cation and voice encryption (CAVE) algorithm to generate a 128-bit subkey called the sharedsecret data (SSD). The SSD has two parts: SSD_A(64-bit) for creating authentication signatures andSSD_B (64-bit) for service providers to allow localauthentication. A fresh SSD can be generatedwhen a mobile returns to the home network orroams to a different system [8].

The mobile uses the SSD_B and the CAVEalgorithm to generate a private long code mask, acellular message encryption algorithm (CMEA)key, and a data key. The private long code maskis used by both the mobile and the network to

change the characteristics of a long code. Thismodified long code is used for voice scrambling,which adds an extra level of privacy over theCDMA air interface. The private long code maskdoes not encrypt information; it simply replacesthe well-known value used to encode a CDMAsignal with a private value known to only themobile and the network. It is therefore extremelydifficult to eavesdrop on conversations withoutknowing the private long code mask.Additionally, the mobile and the network use theCMEA key with the enhanced CMEA algorithmto encrypt signaling messages sent over the airand to decrypt the information received. [8]

A-key security is a critical component of theCDMA system. CDMA allows several A-keydistribution methods to valid users for thepurpose of acquiring subscription-relatedinformation to communicate with the network that is providing service. For alldistribution methods, security data is providedelectronically in an encrypted format. The most secure distribution method uses handsets pre-programmed with the A-key and ESN by the mobile vendor and then assigned a mobile identification number (MIN) by thewireless provider or dealer. This approachensures that neither the equipment manufacturernor the dealer has all three pieces of securityinformation. [7]

CDMA Deployments—Current CDMA-basedWLL networks are deployed in the 450, 800, and 1900 MHz frequency bands. The prominentinfrastructure vendors used in these networksinclude Nortel™, Alcatel-Lucent™, Huawei™,ZTE®, Motorola®, Ericsson™, and Samsung™.

As of September 2006, overall CDMAdeployments include more than 169 operatorswith commercially deployed cdma2000 networksin six continents. An additional 27 operators are scheduled to be deployed in 2007, adding approximately 300 million subscribers [9]. Table 1 provides a brief snapshot of whereCDMA WLL deployments are taking place.

Of the countries listed in Table 1, those thatcontinue to show a lot of activity include Nigeria,Brazil, Iraq, Russia, India, and China. In essence,WLL deployments are taking place in countrieswith little or no existing outside plantinfrastructure (Iraq, Nigeria), large countries withwidely dispersed populations (Brazil, Russia),and countries with large populations and a verylarge subscriber base (India, China).

Through the 1990s,CDMA, and nowcdma2000, hasemerged as the

most widelydeployed WLL

set of standards. As of

September 2006,overall CDMAdeployments

include more than169 operators

with commerciallydeployed

cdma2000networks in

six continents.


WLL Deployment in Developing CountriesIn developing countries, there has always beenstrong interest in using WLL technologies tobridge the “digital divide.” A primary goal ofdeploying WLL systems is to provide universaltelephone access. This is a popular choice indeveloping countries or regions because it is anaffordable alternative to wireline for lowteledensity areas. In most cases, WLL is thequickest and least expensive way to circumventthe dire lack of infrastructure [10].

In developing countries, the driving focus hasbeen to provide basic voice services. Manycountries have focused solely on installing voicenetworks, overlooking the growing use andconvergence of data services. Although voice,and sometimes low-speed data, services are thecurrent technology driver, high-speed data andvideo applications will likely drive the future.

Therefore, data services should be incorporatedfrom the beginning, as is the case with wirelesstechnologies such as cdma2000 1xEV-DO.Incorporating data services should be an integralpart of a developing country’s growth strategy.Data communications should be just as pressing a concern as voice communications in thedeployment of WLL networks [10].

For poorer regions, the Internet presents a uniqueopportunity as well as a unique challenge. Theopportunity derives from new Internet appli-cations of greater value to developing countriesthan to developed ones. Remote regions that havenever been able to afford adequate educationaland health facilities or to attract competentteachers, doctors, agricultural extension officers,and other professionals now have an alternativemedium by which they can benefit from the

services of such professionals, namely, inter-active multimedia. Therefore, deploying datacommunications infrastructure should be regardedas a high priority. In developing regions wherewired infrastructure is scarce, WLL has provideda more-than-viable methodology for doing so. [11]

cdma2000 FOR BROADBAND WIRELESS ACCESS

With the focus shifting from providing basictelephony services in developing countries

or underserved areas to providing more data-centric services, so too is the original ideabehind WLL evolving. Originally posed as “How important is it to provide data services in developing countries?” the idea has nowevolved—as we have seen—to the concept ofproviding ubiquitous fixed and mobile wirelessaccess around the world. No longer is WLLlimited to just voice service in remote regions.Fixed and mobile broadband wireless access is the new “WLL.” That access is the topic of this section.

Examples of Worldwide CDMA Deployments

Eurotel—Czech RepublicIn August 2004, Eurotel, the largest provider ofwireless voice and data services in the CzechRepublic, launched the world’s first commercialCDMA450 1xEV-DO network. The success of itsGSM/general packet radio service (GPRS)solution, along with the growing demand forbroadband data, made Eurotel realize that a high-speed wireless solution was needed. In 2004,mobile penetration in the Czech Republic wasgreater than 98 percent, while high-speed Internetaccess remained less than 3 percent. Recognizinga significant unmet demand for high-speedmobile and fixed Internet access, Eurotel neededto find a quickly deployable technology tocapture the broadband data market whileleveraging its 450 MHz spectrum resources.Furthermore, it needed a technology that couldprovide countrywide coverage cost effectively.The search was narrowed to 1xEV-DO. Thus,CDMA450, combined with the 1xEV-DO capacityand throughput characteristics, was the idealinfrastructure solution for Eurotel’s needs. With224 base stations, Eurotel is now able to providecoverage to 80 percent of the Czech Republicpopulation. Subscribers are able to gain high-speed Internet access without waiting forinstallation. They simply visit their local Eurotelretail outlet and walk home as the owner of awireless access terminal that functions within

Table 1. Where CDMA WLL Deployments Are Taking Place [9]

AFRICA

Algeria, Angola, Congo, Egypt, Ethiopia, Ivory Coast, Kenya, Mali,Mozambique, Namibia, Nigeria, Rwanda, Uganda, Zambia

SOUTH AND CENTRAL AMERICA

Argentina, Brazil, Guatemala, Haiti,Honduras, Peru

MIDDLE EASTIraq, Kuwait, Oman, Pakistan, SaudiArabia, Yemen

CENTRAL AND EASTERN EUROPE

Azerbaijan, Moldova, Poland, Romania,Russia, Ukraine, Uzbekistan

INDIA

CHINA


minutes of purchase. In addition, subscribers can use the service anywhere within the Czech Republic. [12]

Iraq Ministry of CommunicationsThe Iraq Ministry of Communications is movingforward with the development of WLL networksthroughout the country. As announced inSeptember 2006, six licenses have been awardedto provide both national and provincial coverage;this includes an award in 2005 to Iraq’sincumbent telecommunications company, theIraq Telephone and Post Company (ITPC). WLL is not new to Iraq; both ZTE and Huaweihave installed systems in regions throughoutIraq, including in Baghdad and Najaf. ITPC hasindicated that it will use the two Chinese vendors to build out the networks under thelicense it was awarded.

Licensees are authorized to establish operationsto provide local telecommunications services by deploying WLL technologies and to usefrequencies assigned in the 450, 800, and 1900 MHz and 3.5 GHz bands. According tolicense terms, operators have committed to using state-of-the-art technologies for theirdeployments, including 1xEV-DO. In addition,WiMAX will be deployed as both an access and transmission technology. The threenational WLL licensees and three provincial WLL licensees authorized by the NationalCommunications and Media Commission toprovide WLL services in Iraq will join the existing GSM mobile operators as licensedtelecommunications service providers in Iraq. To ensure that Iraq realizes the benefits of having multiple telecommunications networkssimultaneously able to serve users within thesame geographic areas, and as a measure toensure the creation of a distinct class of fullmobility telecommunications service providers,WLL licensees are prohibited from permittingroaming on their networks by other networkoperators or roaming by their users on thenetworks of others. [13]

Sprint™Sprint has been actively expanding, improving,and adding new services to its wirelessbroadband network. This includes plans not only to expand and evolve its high-speed 1xEV-DO network, but also to develop anddeploy the first nationwide WiMAX-basedmobile network. In March 2006, Sprintannounced aggressive plans to expand andevolve Power VisionSM, its 1xEV-DO network,which now covers over half of the US population

with mobile broadband data services [14]. This high-speed wireless network was expectedto reach an estimated 190 million people by theend of 2006. Sprint expected to concurrentlyimplement 3G technology upgrades, known as1xEV-DO Revision A, and anticipates reachingabout 220 million people in the US with thisadvanced network by the end of the third quarterof 2007. Sprint claims to have the most wirelessbroadband coverage of any carrier in the US withaverage download speeds equivalent to DSL(400–700 kpbs). With the evolution to 1xEV-DORevision A, average download speeds improve to 450–800 kbps and average uplink speeds become300–400 kbps (versus the current 70–144 kbps).The 1xEV-DO network was originally launchedin July 2005. [14]

In addition to its 1xEV-DO plans, Sprintannounced in August 2006 its plans to developand deploy the first 4G nationwide broadbandmobile network [15]. The 4G wireless broadbandnetwork will use the mobile IEEE 802.16e-2005WiMAX technology standard. Working withIntel®, Motorola, and Samsung, Sprint willdevelop a nationwide network infrastructure aswell as mobile WiMAX-enabled chipsets that willsupport advanced wireless broadband servicesfor computing, portable multimedia, interactive,and other consumer electronic devices. The Sprint4G mobility network will use the company’sextensive 2.5 GHz spectrum holdings, whichcover 85 percent of the households in the top 100 US markets (the most of any wireless carrierin any single spectrum band). The company’sdeployment plans target a launch of the advancedwireless broadband services in trial markets bythe end of 2007, with plans to deploy a networkthat reaches as many as 100 million people in 2008. [15]

Verizon Wireless®

Verizon Wireless has completed several roundsof successful 1xEV-DO Revision A testing andtrials and, like Sprint, has recently announcedplans to expand its 1xEV-DO network to Revision A. These trials have lead to agreementswith Nortel, Alcatel-Lucent, and Motorola toequip sites with 1xEV-DO Revision A and include core network gear and services [16], [17].Verizon and Motorola have extended theirsupply agreement to include upgrades of existing 1xEV-DO Revision 0 sites and incrementalupgrades of CDMA 1xRTT (1x radio transmissiontechnology) sites to 1xEV-DO Revision A [16].

In August 2005, Verizon and Lucent completedthe industry’s first live, over-the-air calls using

WLL is no longerlimited to

just voice service in remote regions.Fixed and mobile

broadband wirelessaccess is

the new “WLL.”


1xEV-DO Revision A technology; they have sinceconducted live voice over Internet Protocol(VoIP) and video telephony calls using the 1xEV-DO Revision A quality of service (QoS)feature. The increased forward and reverse linkdata speeds reduce data latency and enableoperators such as Verizon to deliver VoIP andother multimedia services [17].

CDMA Performance Data

Expected PerformanceWhat kind of performance should one expectfrom cdma2000 networks? Current CDMAnetwork deployments consist mainly of thecdma2000 1xRTT and 1xEV-DO variety. Withcarriers such as Verizon and Sprint deploying thenext evolution, Revision A, and companies suchas QUALCOMM® and Motorola continuing topush forward with Revisions B and C, CDMAtechnology will continue to be a competitor in thefuture of cellular and wireless broadbanddeployments around the world.

cdma2000 1xRTT supports both voice and data(low-speed data by today’s wireless standards).The voice capacity of 1xRTT typically supports33–35 voice calls per single 1.25 MHz frequencydivision duplex (FDD) channel. The datacapability of a single 1xRTT channel supports bi-directional peak data rates of up to 153 kbpsand an average of 60–100 kbps in commercialnetworks [9]. High-speed data, introduced with1xEV-DO Release 0, provides peak data rates of2.4 Mbps in the forward link and 153 kbps in thereverse link in the same 1.25 MHz CDMA carrier.The average throughput of 1xEV-DO Revision 0is 400–800 kbps in the forward link and 70–100 kbps in the reverse link. [9] Despite thefact that Revision 0 no longer supports circuit-switched voice traffic, the migration path forCDMA relies on increased performance and datarates, particularly in the reverse link, and on theevolution of providing voice service through apacket-switched data network.

With cdma2000 1xEV-DO Revision A technology, peak download data rates increase to 3.1 Mbps and peak upload data rates increase to 1.8 Mbps. As stated earlier,average download speeds improve to 450–800 kbps, and average uplink speeds become 300–400 kbps [14]. Revision A starts to make practical the shift to an all-IP-basednetwork for voice, data, and other multimedia-based services. However, a long-term strategy is necessary for continued evolution andcompetitive positioning.

cdma2000 1xEV-DO Release B, also called multi-carrier EV-DO, introduces a 64 quadratureamplitude modulation (QAM) scheme anddelivers peak rates of 73.5 Mbps in the forwardlink and 27 Mbps in the reverse link byaggregating fifteen 1.25 MHz carriers within 20 MHz of bandwidth. A single 1.25 MHz carrier and an aggregated 5 MHz carrier in theforward link deliver a peak rate of up to 4.9 Mbpsand 14.7 Mbps, respectively. Revision B will becommercially available in 2008 [9].

Actual PerformanceNow that the expected performances for thedifferent flavors of cdma2000 are known, whatare users actually experiencing? With Revision Ajust beginning to be deployed and Revision B stillsome time out, actual performance data focuseson 1xRTT and 1xEV-DO Revision 0.

One snapshot of typical user experience wasobtained from tests performed by PC World.These tests characterized Verizon’s performanceas of January 2006 in various locations and usagepatterns. Reported results show typical 1xEV-DOdata speeds of 300–500 kbps, with bursts of 2.4 Mbps, and 1xRTT data speeds (in areas where1xEV-DO was not available) of approximately 70 kbps. [18]

Additional information from a popular onlineand mobile phone community, Howard Forums,suggests that average performance on Sprint andVerizon 1xEV-DO services, as of February 2006, is774 kbps for Sprint and 594 kbps for Verizon,with respective peak rates of 1.87 Mbps and 1.98 Mbps. These values are averages for over 100 tests on each network for users throughoutthe country [19]. Additional Howard Forumsstatistics for these tests are shown in Table 2.

Test results obtained from Huawei cdma2000network equipment installed in the BechtelTelecommunications Laboratory demonstratesimilar findings. The tests used file transferprotocol (FTP) on a local network with both

Table 2. User Performance for Sprint Nextel and VerizonWireless 1xEV-DO Networks as of February 2006 [19]

SprintNextel

VerizonWireless

No. of Entries 122 159

Avg. Forward Link Throughput 774 kbps 594 kbps

Median 764 kbps 610 kbps

Peak 1.868 Mbps 1.976 Mbps

% > 1 Mbps 27% 7%

% > 600 kbps 64% 51%

% < 400 kbps 17% 24%

For serviceoperators, there are

three primary VoIP drivers:

network efficiency,more efficientspectrum use, and the ability

to enhance voice service

portfolios.


1xRTT and 1xEV-DO networks. Since the testswere conducted in a laboratory environment andwith a lightly loaded system, these values couldbe assumed as generally a best-case scenario.

For cdma2000 1xRTT, the single-user averagedownload speed is 126 kbps and average uploadspeed is 100 kbps. For cdma2000 1xEV-DO, thesingle-user average download speed is 872 kbpsand average upload speed is 141 kbps. Whenadditional users are added, the averagethroughput per user goes down. However, thebursty nature of data causes the combinedrealized throughput of the single cdma2000carrier to appear to be much larger. The results of these tests are shown in Table 3.

1xEV-DO Revision A and the Introduction of VoIP

VoIP DriversBecause 1xEV-DO is an inherently all-IP-basedarchitecture, it does not support circuit-switched

voice. Therefore, carriers that want to supportboth voice and high-speed data must deploy ahybrid cdma2000 1xRTT and 1xEV-DO network.This approach, although used by CDMA carrierstoday, is obviously not an ideal solution movingforward. To be supported by an all-IP-basedarchitecture such as 1xEV-DO, voice servicesmust migrate from conventional circuit-switchedto packet-switched. As can be seen from even thesimple performance data given above, to thispoint it has not been possible to support large-scale VoIP services on existing 1xEV-DOnetworks. However, with the introduction of1xEV-DO Revision A and future revisions, thepotential for VoIP and other multimedia servicescan be realized.

For service operators, there are three primaryVoIP drivers: network efficiency, more efficientspectrum use, and the ability to enhance voiceservice portfolios [20]. VoIP over 1xEV-DObenefits from several advantages. First andforemost, there is only one network to supportand maintain. Another important benefit involvesspectrum use. As shown in Figure 4, currentCDMA networks require two radio channels: one for 1xRTT voice and one for broadband data.This separation of voice and data results ininefficient spectrum use because it is not possibleto offload traffic from one channel to the other tobalance the load. Migrating to mobile VoIP overEV-DO allows voice and broadband data to besimultaneously carried over the same channel or

Table 3. cdma2000 Test Results

1xRTT Downlink (kbps) Uplink (kbps)

1 User – 1 Terminal 126 100

4 Users – 4 Terminals 326 (82/user) 320 (80/user)

1xEV-DO Downlink (kbps) Uplink (kbps)

1 User – 1 Terminal 872 141

4 Users – 1 Terminal 733 (183/user) 175 (44/user)

4 Users – 4 Terminals 1,361 (340/user) 458 (114/user)

CircuitCore

PTSN

MSC

PDSN

BSC

BTS

VoiceData

BTS

1xRTTCircuit Voice Network

EV-DO Release 0Packet Data Network

BSC

PacketCore

PDSN

Figure 4. Traditional cdma2000 1xRTT/1xEV-DO Voice and Data Network [21]

Each evolutionarystep of cdma2000

builds on theinherent advantages

of CDMAtechnologies and

introducesenhancements that

further increasespectral efficiencies

and datathroughput, while

supporting theconvergence of

fixed, mobile, andmulticasting

networks throughall-IP delivery.


carrier in a load balancing configuration. MobileVoIP gives operators efficiency and cost savingsthrough a single packet-based core network [21].A flatter network architecture is realized, asshown in Figure 5.

1xEV-DO Revision A AdvancementsNot only can VoIP replace traditional voiceservices, it can also make new services possiblebecause both voice and data communicationstravel over the same network. Examples includepush-to-talk (PTT), video telephony, click-to-dial,whiteboarding, and advanced accessibilityfeatures. Although services such as PTT are notnew, the increased performance realized inRevision A enables improved implementationcompared with what can be done on currentnetworks. As proof of this, Sprint has chosen1xEV-DO Revision A as the technology it plans to use to migrate millions of integrated digitalenhanced network (iDEN) subscribers to a morescalable version of PTT based on IP. Sprint’soffering will use an optimized version of PTT, QUALCOMM’s QChat™, with very fast callset-up times. Coupled with QChat, Revision Awill underpin the first cellular network to deliversub-second PTT call set-up [21].

1xEV-DO Revision A adds several advancementsfor VoIP and other multimedia traffic. Theseinclude increased data rates, end-to-end QoS,support for short and multiuser packets, andsupport for header compression. Compared withRevision 0, Revision A increases the peak forward

link data rate to 3.1 Mbps and, more importantly,increases the peak reverse link data rate to 1.8 Mbps. This dramatically enhanced uplinkdata rate enables Revision A networks to supportsignificantly more voice communications than ispossible under Revision 0.

The 1xEV-DO forward link uses TDM to sendpackets to various users; this requires ascheduling function to select which user shouldgain access to the air link at any given time. 1xEV-DO Revision A adds the ability for theforward link scheduler to coordinate the use ofthe air link with the various devices using it. On the reverse link, Revision A permits the use of higher power for QoS packets in order toreduce the number of transmissions and retriesnecessary to successfully send these packets [22].

1xEV-DO Revision A supports a wide variety ofshorter packet transmissions. This is necessarybecause, unlike many data applications, voiceapplications regularly transmit relatively shortpackets. The shorter packets can be transmitted inless time and allow more users to access thenetwork with low latency. In addition, 1xEV-DORevision A includes support for multiuserpackets. This capability allows a long physicallayer packet to be addressed to separate users,again reducing air link overhead as well as per-user delay.

Along the same lines as short and multi-userpackets, 1xEV-DO Revision A supports header

PDSN

MediaGateway

SIPServer

VoIPData

BTS

EV-DO Revision APacket Voice and Data Network

BSC

PacketCore

PSTN

PDSN

Figure 5. cdma2000 1xEV-DO Revision A Voice and Data Network [21]

cdma2000 1xEV-DO Revision A

adds severaladvancements for

VoIP and othermultimedia traffic.

These includeincreased data

rates, end-to-endQoS, support for

short and multiuserpackets, and

support for headercompression.


compression. The overhead required in currentnetworks results in an inefficient overhead-to-payload ratio. Typical ratios can be on the orderof 2:1, or a required header twice as large as theactual voice payload. Such inefficiencies affect theoverall network capacity to handle voice traffic.Header compression reduces these headers fromas much as 40 bytes down to approximately 2 bytes. [20, 22]

Putting VoIP and 1xEV-DO Revision A TogetherAccording to simulations performed in [22], VoIPover 1xEV-DO Revision A supports 40–60simultaneous users. Even with a guaranteedvoice quality of 270 ms mouth-to-ear delay for 40–60 simultaneous VoIP users, 1xEV-DORevision A makes 400 kbps per sector availableon the forward link for data users. This is in sharp contrast to 1xRTT, which cannot deliver simultaneous voice and broadband data under any circumstances. Figure 6 shows the relationship between the number of voice subscribers supported and the offered data capacity.

In June 2006, QUALCOMM successfullydemonstrated the potential for using VoIP across1xEV-DO Release A. The field tests, conductedover a cdma2000 1xEV-DO Revision A system inone sector within a single 1.25 MHz channel in afully mobile configuration, showed a voicecapacity capable of supporting 62 simultaneous

calls. These field tests validate the quality andcapacity of fully mobile VoIP over 1xEV-DORevision A and pave the way to large-scalecommercial trials by network operators [23].

The Path of cdma2000 EvolutionThe principle underlying the path of cdma2000evolution is backward and forward compatibility.This compatibility allows operators to seamlesslyand cost-effectively upgrade their existingnetworks to enhance capabilities and advanceservices. Each evolutionary step of cdma2000builds on the inherent advantages of CDMAtechnologies and introduces enhancements thatfurther increase spectral efficiencies and datathroughput, while supporting the convergence of fixed, mobile, and multicasting networksthrough all-IP delivery. [9]

Evolutionary Dead EndsIn contrast to previous CDMA evolutionstrategies that planned on cdma2000 1xEV-DV(evolution-data/voice) and cdma2000 3x (multi-carrier CDMA using three 1.25 MHz carriers), thecurrent evolution strategy is focused on upgradesto the existing 1xEV-DO technology path. In thismanner, the intent of both 3x and 1xEV-DV will be embraced in the advanced revisions of1xEV-DO. Since 1xEV-DO was originally thoughtof as a data-only service, cdma2000 1xEV-DV was envisioned as a technology to provide both

Voice Users

Sect

or Th

roug

hput

(kbp

s)

1,400

1,200

1,000

800

600

400

200

0 5 10 15 20 25 30 35 40 45 50

DataThroughput

with 40Simultaneous

VoIP Calls

Figure 6. Simultaneous VoIP and Data Sector Throughput [21]

cdma2000 1xEV-DO Revision B

introduces a 64 QAM modulationscheme and delivers

peak rates of 73.5 Mbps in theforward link and 27 Mbps in thereverse link by

aggregating fifteen1.25 MHz carriers

within 20 MHz of bandwidth.


high-speed data and voice services using thesame 1.25 MHz CDMA carrier. However, 1xEV-DORevision A technology incorporates many of theadvances originally planned for 1xEV-DV. Thesame is true for cdma2000 3x. Future 1xEV-DORevisions B and C are expected to providescalable bandwidths by making use of multiplecdma2000 EV-DO carriers. The present cdma2000migration path, as outlined by the CDMADevelopment Group (CDG), is shown in Figure 7.

1xEV-DO Revision Acdma2000 1xEV-DO Revision A is an all-IP-basedair interface that enables the integration of VoIP, high-speed packet data, and enhancedmultimedia capabilities. As previously discussed,1xEV-DO Revision A represents a major step inthe evolution of cdma2000 standards towardconverged communication networks andubiquitous delivery of voice and data servicesacross fixed and wireless networks. The majorenhancements of Revision A include increaseddata throughputs, increased rate quantization onthe forward and reverse links, reduced latency,and optimized QoS. In addition, an orthogonalfrequency division multiplexed (OFDM)waveform is introduced to offer high capacitymulticast capabilities that allow operators to offerlower cost multicast services while maintaining a robust high-speed mobile network withcdma2000. [24]

1xEV-DO Revision Bcdma2000 1xEV-DO Revision B, also known asmulticarrier CDMA, introduces a 64 QAMmodulation scheme and delivers peak rates of73.5 Mbps in the forward link and 27 Mbps in thereverse link by aggregating fifteen 1.25 MHzcarriers within 20 MHz of bandwidth. However,the aggregation of three channels within 5 MHzfor 9.3 Mbps on the forward link and 5.4 Mbps onthe reverse link is a more likely configuration incommercial networks [11]. This scalable approachto allocating bandwidth allows a linear approachto aggregating carriers, depending on networkdemands and spectrum availability. Furtherefficiencies can be realized through dynamic loadbalancing, or allocating radio spectrum forparticular locations, customers, or applicationsthat require increased data throughputs. Any 1.25 MHz carrier or group of carriers can be usedas needed [25].

In addition to supporting mobile broadband dataand OFDM-based multicasting, as introduced in1xEV-DO Revision A, the lower latencycharacteristic of 1xEV-DO Revision B furtherimproves the performance of delay-sensitiveapplications such as VoIP, PTT over cellular,video telephony, and the like. 1xEV-DO Revision B systems, which will be commerciallyavailable in 2008, maintain backward andforward capability with previous revisions of

cdma20001x

cdma20001xEV-DO

1xEV-DORevision A

(Note 1)

VolP

CDMA CDMA/TDM CDMA/OFDM CDMA/OFDMA/MIMO/SDMA

cdma2000 Path (1.25 MHz Channel)

2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

DL: 3.1 MbpsUL: 1.8 Mbps

DL: 2.4 MbpsUL: 153 kbps

DL: 153 kbpsUL: 153 kbps

DL: 3.1–73 MbpsUL: 1.8–27 Mbps

1.25–20 MHz

Requirements:DL: 70–200 MbpsUL: 30–45 Mbps

1.25–20 MHz

1xEV-DORevision B(Notes 1, 2)

1xEV-DORevision C

(Note 3)

NOTES: Timeline depicts initial commercial availability of each technology. Those introduced beyond 2008 are under standardization and are subject to variability. 1. EV-DO Revision A and Revision B incorporate OFDM for multicasting.2. Data rates of 73 Mbps for the DL and 27 Mbps for the UL figures are based on a 2 x 20 MHz allocation.3. Data rate is dependent on level of mobility.

Figure 7. cdma2000 Migration Path [9]

cdma2000 1xEV-DO Revision C will use technologies

such as OFDMA,SDMA, and MIMO

to achieve higher peak data rates,

extremely lowlatency, and very

high spectralefficiency.


1xEV-DO. [25] Figure 8 shows the expectedincrease in performance for 3-channel 1xEV-DORevision B, compared with previous versions of1xEV-DO and 1xRTT.

1xEV-DO Revision CThe latest enhancement in the evolution ofcdma2000 1xEV-DO is Revision C. According to a system requirements document for the next air interface evolution of cdma2000 approved inMay 2006, the forward link and reverse link peakdata rates, using scalable bandwidths up to 20 MHz, should be respectively capable of up to500 Mbps and 150 Mbps in a stationary indoorenvironment and up to 100 Mbps and 50 Mbps in a mobile environment [26]. Building onRevision B, the technology has a flexible anddynamic mode of operation to combine andallocate spectrum as needed to support a largevariety of user applications and activities. [25]

Revision C is planned for commercial availabilityin 2008 and uses technologies such as orthogonalfrequency division multiple access (OFDMA);spatial division multiple access (SDMA); andmultiple-input, multiple-output (MIMO) toachieve higher peak data rates, extremely lowlatency, and very high spectral efficiency. The framework employs OFDMA on the forwardlink and supports several advanced antennatechniques, including MIMO and SDMA. The reverse link employs quasi-orthogonal

transmissions based on OFDMA, along with non-orthogonal user multiplexing with layeredsuperposed OFDMA (LS-OFDMA). The reverselink also supports CDMA transmissions forcontrol and for low-rate, low-latency traffic.Detailed technical specifications are expected tobe completed by early in the second quarter of 2007. [26]

CONCLUSIONS—MOVING FORWARD

This paper takes a very wide look at thedevelopment and implementation of WLL

systems and in particular the predominanttechnology employed in these systems,cdma2000. Despite the fact that WLL systemshave mostly been used to provide basiccommunication services, either as an alternativecarrier or in remote areas where legacy servicedoes not exist, the growing demand for high-speed data services, along with advances intechnologies, has continued to morph thelandscape. Data services are no longer an option,but rather the foundation for all other services.Therefore, the shift has been made from WLLsystems to a more generic form of broadbandwireless access.

This paper focuses on cdma2000. However, as current 3G cellular networks such asUMTS/high-speed packet access (HSPA) and

0.153 1.8 3.1 5.4 9.3

3xMC-DO5 MHz

1xEV-DORevision A

1xEV-DORevision 0

1x

3xMC-DO5 MHz

Mbps

9.3

Forward Link Performance Improvement Comparative Downloads

3.1

2.4

0.153

0.2

0.5

0.7

10

3

9

12

183

8.7

26

33

523

200 KBPicture

3.5 MBMP3File

10 MBPowerPoint®

Presentation

1xEV-DORevision A

1xEV-DORevision 0

1x

SecondsMbps

Figure 8. Comparative Performance of 1xEV-DO Revision B [25]

As current 3Gcellular networksevolve and new

networks such asWiMAX are

deployed, thedistinction betweenthem continues toblur because the

technological basesof IEEE 802.16e,LTE, and EV-DOrevisions look

increasingly similar.


cdma2000 1xEV-DO evolve (UMTS to HSPA and eventually long-term evolution [LTE], and1xEV-DO from Revision A to Revision B and on to Revision C) and new networks such asWiMAX are deployed, the distinction betweenthem continues to blur because the technologicalbases of IEEE 802.16e, LTE, and EV-DO revisions look increasingly similar. Thetechnological similarities include such things asan OFDM basis, smart antenna techniques, an all-IP network, and non-hierarchical network structures. [27]

Even so, the importance of each system willultimately be decided by commercial and political factors, not technological ones. Because of this, cdma2000 1xEV-DO will have an important place alongside other systems in the mobile broadband picture moving forward. [27]

TRADEMARKS

Alcatel-Lucent is a trademark of Alcatel-Lucent.

cdma2000 is a registered trademark andcertification mark of the TelecommunicationsIndustry Association (TIA-USA).

Ericsson is the trademark or registered trademark of Telefonaktiebolaget LM Ericsson.

Huawei is a trademark of Huawei TechnologiesCo., Ltd.

Intel is a registered trademark of IntelCorporation in the United States and othercountries.

Motorola is registered in the U.S. Patent andTrademark Office by Motorola, Inc.

Nortel is a trademark of Nortel NetworksLimited.

PowerPoint is a registered trademark ofMicrosoft Corporation in the United States and other countries.

Power Vision is a service mark and Sprintis a trademark of Sprint Nextel.

QChat is a trademark of QUALCOMMIncorporated.

QUALCOMM is a registered trademark ofQUALCOMM Incorporated.

Samsung is a trademark of Samsung in theUnited States or other countries.

Verizon Wireless is a registered trademark ofVerizon Trademark Services LLC or its affiliatesin the United States and/or other countries.

WiMAX and WiMAX Forum are trademarks ofthe WiMAX Forum.

ZTE is a registered trademark of ZTECorporation.

REFERENCES

[1] “Wireless Local Loop (WLL),” IEC white paper(http://www.iec.org/online/tutorials/wll).

[2] V.K. Garg and E.L. Sneed, “Digital Wireless LocalLoop System,” IEEE Communications Magazine,Vol. 34, No. 10, October 1996, pp. 112–115.

[3] W.C.Y. Lee, “Spectrum and Technology of aWireless Local Loop System,” IEEE PersonalCommunications, Vol. 5, No. 1, February 1998, pp. 49–54.

[4] M. Naidu, “CDMA2000 for Wireless in LocalLoop Networks,” QUALCOMM white paper,December 2004.

[5] J. Cawley, Inquam, “Regulatory Policies andSpectrum Requirements for use of CDMA in the 450 MHz Band,” presentation at CDMA450Seminar, November 2004.

[6] “CDMA Enabling WLL Services and Bridging the Digital Divide,” QUALCOMM presentation,December 2002.

[7] “CDMA End-to-End Security,” Nortel Networks,September 2004.

[8] C. Wingert and M. Naidu, QUALCOMM,“CDMA 1xRTT Security Overview,” August 2002.

[9] CDMA Development Group (http://www.cdg.org and http://www.cdg.org/technology/3g/evolution.asp).

[10] M. Kibati and D. Krairit, “Wireless Local Loop in Developing Regions: Is It Too Soon for Data?”Communications of the ACM, Vol. 42, No. 6, June 1999, pp. 60–66.

[11] CDMA Development Group, “Taking CDMA2000into the Next Decade,” October 2005.

[12] “Success Story: Eurotel,” Nortel case study, 2005.[13] Iraqi National Communication and Media

Commission, Official Announcement Award of Wireless Local Loop Licenses, September 2006(http://www.ncmc-iraq.com/wllAnnouncement.htm).

[14] “Sprint Extends Mobility Leadership withAggressive Broadband Network Expansion,”Sprint news release, March 2006 (http://www2.sprint.com/mr/news_dtl.do?id=11040).

[15] “Sprint Nextel Announces 4G WirelessBroadband Initiative with Intel, Motorola andSamsung,” Sprint news release, August 2006(http://www2.sprint.com/mr/news_dtl.do?id=12960).

[16] “Verizon Wireless and Motorola Agree to Expand and Upgrade CDMA 1x Sites to EV-DORev. A,” Verizon news release, September 2006(http://news.vzw.com/news/2006/09/pr2006-09-12.html).


[17] “Verizon Wireless Selects Lucent Technologies forCDMA2000 1xEV-DO Revision A Technology,”Verizon news release, June 2006(http://news.vzw.com/news/2006/06/pr2006-06-27.html).

[18] C. Null, “Broadband to Go,” PC World magazine, January 2006 (http://www.pcworld.com).

[19] Howard Forums, February 2006 (http://www.howardforums.com/showthread.php?threadid=846131).

[20] Airvana VoIP Technology (http://www.airvananet.com/print/technology_voip.htm).

[21] P. Callahan, “Mobile VoIP over 1xEV-DO,”Airvana technical white paper, July 2006.

[22] M. Yavuz, S. Diaz, R. Kapoor, M. Grob, P. Black,Y. Tokgoz, and C. Lott, “VoIP over CDMA20001xEV-DO Revision A,” IEEE CommunicationsMagazine, Vol. 44, No. 2, February 2006, pp. 88–95.

[23] “QUALCOMM Successfully Demonstrates FullyMobile VoIP Calls Across a Number of Field Test Environments,” QUALCOMM press release,June 2006.

[24] CDMA Development Group, “CDMA2000 1xEV-DO Revision A, The Gateway to TrueMobile Broadband Multimedia,” August 2006.

[25] “CDMA2000 EV-DO Revision B,” Motorola white paper, May 2006.

[26] “CDMA Industry is Ready to Meet Future Market Needs Head On,” 3GPP3 news release,August 2006 (http://www.3gpp2.org/Public_html/News/Release_15Aug2006.pdf).

[27] E. Weinman, “WiMAX: State of the Industry,”WiMAX Business & Technology Strategies, October 2006.

BIOGRAPHYNathan Youell joined Bechtel in2001 and is currently a systemsengineer with the BechtelFederal Telecoms StrategicInfrastructure Group. He is theresident subject matter expertfor wireless systems and isresponsible for testing andevaluating telecommunicationsequipment, as well as modeling

and simulating critical infrastructure, with a primaryfocus on telecommunications systems.

Previously, as a staff scientist/engineer and themanager of the Bechtel TelecommunicationsLaboratory, Nathan gained and then provided expertise in developing and implementing test plansand procedures. He was instrumental in creating the Bechtel Training, Demonstration and Research(TDR) Laboratory in Frederick, Maryland, and theBechtel Wireless Test Bed (BWTB) in Idaho Falls, Idaho. He also tested numerous telecommunicationsequipment and technologies, including TMA, 802.11,802.16, GSM, DAS, DWDM, FSO, microwave andmillimeter wave radio, and wireless repeater. Earlier,Nathan was an RF engineer in the New York andWashington, DC, markets as part of Bechtel’snationwide build-out contract with AT&T Wireless.

Nathan received both his MS and BS degrees in Electrical Engineering from Clemson University,South Carolina.

Bechtel Telecommunications, a unit of BechtelCorporation, provides turnkey deployment services

that include network planning, RF design, engineering,and project and construction management for thedeployment of wireless, wireline, and othertelecommunication facilities worldwide. We areheadquartered near Washington, DC, and supported byour global network of offices.

With more than 100,000 new build, overlay, and upgrade wireless sites deployed; 23,000 kilometers ofwireline fiber laid; and numerous communicationcenters constructed around the globe, Bechtel is theglobal company of choice for telecommunicationsinfrastructure deployment.

Bechtel has continually set the bar for telecom-munications network project deployment for clientsthroughout the world. Due to our success with tightschedules and cost effective delivery, we are currentlymanaging network expansions and upgrades for majoroperators in the US, Europe, and the Asia-Pacific region.Signature record-breaking projects include Cingular,AT&T Wireless, Vodafone, Verizon, Metromedia FiberNetworks, XO, WINFirst, Viatel, Pangea, and Equinix.

BECHTEL TELECOMMUNICATIONS

Telecommunications Leadership

TIMOTHY D. STATTONExecutive Vice President and Director, Bechtel Group, Inc.President, Bechtel Telecommunications

MIKE HICKEY Principal Vice President and Manager of Functional Operations

JAMES A. IVANYPrincipal Vice President and Business Manager and Chief Financial Officer

J. S. (JAKE) MACLEODPrincipal Vice President and Chief Technology Officer

LEE LUSHBAUGH Principal Vice President and General Manager, Americas

RICK ASTLEFORDPrincipal Vice President andGeneral Manager, Europe, Africa, Middle East, Asia, and Asia-Pacific


Bechtel Corporation is a privately heldcompany headquartered in San Francisco,

California, with approximately 40,000 employees,more than 40 offices worldwide, and 2005revenues of approximately $18.1 billion. Foundedin 1898, Bechtel has been under the leadership ofits founding family for four generations. Riley P.Bechtel, great-grandson of the founder, is thecurrent chairman and chief executive officer.Bechtel is one of the world's premier engineering,construction, and project management companies.

• The Financial Times recently listed Bechtelas the “World’s Most Respected Company”in the property and construction industries.

• Global Finance magazine named Bechtel the “World’s Best Company” in theconstruction sector for 5 years in a row.

• Engineering News-Record ranked Bechtel asNumber 1 on the list of the “Top 400Contractors” for 7 consecutive years.

• Engineering News-Record also rankedBechtel Number 1 TelecommunicationsContractor from 2002 to 2006.

Because of our unparalleled industry recognitionand financial stability, we are the partner of choicefor network deployment.

Bechtel Corporation

TECHNOLOGY EVALUATION

Key to Bechtel's deployment success is ourtechnology expertise. To provide our

clients with complete turnkey deploymentservices, we offer expert technology researchand testing services. We have assembled afull-service Technology Group of leadingprofessionals that provides our customerswith neutral technology services andoversees our in-house telecommunicationslaboratories: the Bechtel Wireless Test Bed(BWTB) and the Training, Demonstration and Research (TDR) Laboratory.

The BWTB, an over-the-air wireless test bed, enables us to analyze network products and technologies for our clients in a real-world environment. It can bedynamically configured to meet specifictesting requirements without impacting the client’s network.

The TDR Laboratory offers our clients the ability to integrate equipment frommultiple vendors to identify interoperabilityissues, monitor product performance, andensure compatibility with legacy systems,thereby reducing the risk of imple-menting new technologies and paving the way for investment in network upgrades or expansion.

92

bttjv5n1

Documents