PIMRC and Mobile Communications The Next 25 Years

Dr. Kenneth Stewart, Intel PIMRC 2014 4 September 2014 Washington DC

The opinions expressed are those of Dr. Kenneth Stewart and may not reflect the opinions, conclusions or technical position of Intel Corp. or its affiliates. *Other names and brands may be claimed as the property of others DO NOT FORWARD

“I always avoid prophesying beforehand because it is much better to prophesy after the event has already taken place”

Winston Churchill

PIMRC 2014 2

In The Beginning… March 1983: Motorola DynaTAC 8000x*

Motorola DynaTAC 8000X DynaTAC 'Dynamic Adaptive Total Area Coverage' * Introduced: March 1983 Talk Time: 30 minutes Stand By Time: 8 hours Recharge Time: 10 hours Dimensions: 33 x 4.5 x 8.9cm Volume: ~1300cc Display: LED Memory: 30 Number, <1MB (256B Pages) Air Interface: AMPS (30kHz FM) Price: $3,995 (~$9,400 2014 $’s)

Motorola DynaTAC 8000X

VW Beetle 1966 1300cc Engine

PIMRC 2014 3

Then and Now Motorola DynaTAC 8000*x vs. Intel Merrifield Motorola DynaTAC Introduced: March 1983 Weight: 790g Dimensions: 33 x 4.5 x 8.9cm Volume: ~1300cc Display: LED Memory: <1MB (256B Pages) Air Interface: AMPS Bands: 850MHz

Intel Merrifield (Atom Z34xx) Introduced: Q1-2014 Weight: 160g (5x) Dimensions: 6.5 x 13.2 x 0.88cm Volume: ~75cc (17x) Display: 1280x720, 326 dpi (NaNx) Memory: 1GB LPDDR3 RAM (NaNx) Air Interface: 2G/3G/4G, .11ac, BT/LE, GNSS (NaNx) Bands: 2G=4, 3G=5, 4G=9 (18x)

Data Rate = 10kbps (Codec Equivalent)

Data Rate = 150Mbps (Assuming LTE Cat. 4 XMMTM7160)

Wide Area Data Rate Doubling Every 2 Years 1983-2014 Comparable to Moore’s Law

PIMRC 2014 4

Radio Access Network Evolution 1989-2014

OFDM/S-OFDMAFDD, TDD, HD-FDDTDMA – HD-FDD

1989 - GSM 2014 - LTE

Sectored Re-Use 1/21 –> 3/7 -> Fractional

Full Re-UseMU-MIMO >1

Convolutional, K=5 PCCC, K=9

None 8-StreamClosed & Open Loop

None Joint MIMO

None LTE-WiFi Multi-Flow Aggregation

MultipleAccess

FrequencyRe-Use

Forward Error Correction

Spatial Multiplexing

Site/AP Coordination

Multi-RAT Operation

MLSEReceiver Design Quasi-ML &Interference Suppression

PIMRC 2014 5

Network Evolution Today’s Heterogeneous Network

Small Base Stations (SBS)LTE, HSPA+, WiFi

Macro SiteLTE, HSPA+

Metro Fiber Ethernet Ring

Macro SiteLTE, HSPA+

Macro SiteLTE, HSPA+

VDSL or VDSL2Copper backhaul

Local Fiber Ethernet Ring

mm-Wave BackhaulFixed mm-wave point-point links, 28/33GHz Heterogeneous Network

“HetNet” RAN

Centralised Multi-cell Scheduling

Slow inter-cell Radio Resource ManagementSBS

LTE, HSPA+, WiFi

IuPS

IuCS

Iub

PSTN

Internet

IMSS1-U

S1-MME

S11

UMTS Data – HSPA, HSPA+

WCDMA Voice

LTE Data + Voice

MME

S-GW PDN-GW

GGSNSGSNRNC

MSCVLR

MGW

Operator Core Network(LTE, WCDMA Example)

C/U-Plane P-CellU-Plane S-Cell

Aggregated Macroand Small Cell

Macro-Small CellOverlay/Underlay

10's of small cells per macro-cell

Centralised Multi-cell Scheduling

Fast inter-cell RRM and interference coordination.

Multi-Flow Channel Concatenation

Multi-site traffic channel combination

Macro SiteLTE, HSPA+

PIMRC 2014 6

Network Evolution Virtualization and Coordinated or Distributed RAN

Multi-RAT Baseband Unit (BBU)Multi-RAT: LTE (FDD/TDD), HSPA+

Baseband signal processing resource shared between 10's or 100's of logical cells

Distributed (‘C’) Radio Access Network (‘C’-RAN)

Centralised SchedulerFast scheduler, inc.

interference coordinationCPRI Fiber Distribution

Dark or operator fiber – Multi- λ CWDM/DWDM, Tbps Backhaul

Multi-Band Remote Radio Head (RRH)

Full-band, multi-band operation, arbitrary waveform transmission

Band 1

Band 3

WCDMA/HSPA+ SISO

CDMA 1xRTT, 1xEVDO SISO

LTE MIMO

RAT-Specific Site Re-use

Star Topology

Cascade Topology

IuPS

IuCS

Iub

PSTN

Internet

IMSS1-U

S1-MME

S11

UMTS Data – HSPA, HSPA+

WCDMA Voice

LTE Data + Voice

MME

S-GW PDN-GW

GGSNSGSNRNC

MSCVLR

MGW

Virtualized Core Network(LTE, WCDMA Example)

Processing HubEnterprise core, central office, stadium or venue

Band 1

Band 3

Band 2

PIMRC 2014 7

Wide Area Network (WAN) Evolution 2G to 4G: GSM/EDGE, UMTS/WCDMA, CDMA, LTE

PIMRC 2014 8

Mobile Communications Economic Impact Examples Since PIMRC Inception

1. Source: CTIA, Chetan Sharma Consulting 2. Source: GSM Association 3. Source: International Monetary Fund 4. ABI Research

Example – U.S. Subscriber Growth Economic Indicators

0

50

100

150

200

250

300

350

1990 1995 2000 2005 2010 2015 2020

U.S. Cellular Connection Growth1

(Millions)

U.S. Cellular ConnectionsU.S

. Cel

lula

r Con

nect

ions

Year

Global Scale Global Mobile Subscriptions (2014): 7.5BnWorld Population (2014): 7.2Bn

Economic Impact

Mobile Industry Value (2013): $2.3Tn2

CountryUSAChinaJapanGermanyFranceUKBrazilRussia

GDP ($Tn)3

16.89.24.93.62.72.52.22.1

Technical Breadth

Device Shipments (2013)4

WiFi Enabled: 1.2BnBluetooth Enabled: 1.5BnGlobal WiFi Hotspots (2013): 4.2M

PIMRC 2014 9

Semiconductor Innovation and Moore’s Law Future Viability – Transistor Cost and Area Density

PIMRC 2014 10

Semiconductor Architectural Innovation New Topologies

2D Integration 3D Integration

Logic Memory Power Reg. Radio Sensors Photonics

Heterogeneous System Integration1

1. IEDM 2011, The Evolution of Scaling from the Homogeneous Era to the Heterogeneous Era, M. Bohr 2. IEDM 2012, Uniform Methodology for Benchmarking Beyond-CMOS Logic Devices, D. Nikonov, I. Young 3. IEDM 2012, The Ultimate CMOS Device and Beyond, K. Kuhn et al.

Beyond CMOS Devices2,3

Advanced Interconnect

Cu Wires at 17nm. Colour indicates crystal orientation.

Planar FinFET

PIMRC 2014 11

Processor Core Energy Efficiency Generational Evolution

PIMRC 2014 12

Network Design Evolution Key Enablers for Next-Generation Network Performance

System Densification

Advanced Source Coding

Enhanced Spectral Efficiency

New Spectrum Hz

nodes/km2

bps/Hz/node

bps/km2Network Performance

• Small Cells, Radio Heads, HetNet• NNT – New Network Topologies• AIS – Advanced Interference Suppression

• Sub-6GHz and mm-Wave• LSA/SAS, White-Space, Unlicensed• Inter-RAT Aggregation – WiFi, WiGig, LTE

• H.265, EVS, Optimal Scheduling• MMIMO and Non-Orthogonal Modulation• Next-generation Codes• Multi-Flow Node Aggregation

Hz nodes/km2 bps/Hz/node bps/km2

Fundamental Network KPI = Capacity per Unit Area

PIMRC 2014 13

Adding to the Spectrum Pool New Bands, Flexible Sharing

PIMRC 2014 14

Advanced Receiver Evolution Interference Management, HetNets and Higher Frequency Re-use

Note 1: Multiple non-standardized and proprietary techniques have been deployed in commercial implementations. Note 2: RAT – Radio Access Technology Note 3: SIC - Successive Interference Cancellation

Year Standard Technique1 RAT2 Description

2003 SAIC GSM Single Antenna Interference Cancellation (SAIC), single interferer suppressed by exploiting GMSK constellation redundancy.

2003 UMTS Type 2 HSPA Single antenna equalizer replaces RAKE receiver.

2006 DARP GSM Extends SAIC to dual-antennas (“Rx diversity”) at the device.

2006 UMTS Type 2i HSPA Single antenna equalizer with dominant interferer interference suppression (usually single cell), usually SIC3 architecture.

2007 UMTS Type 3i HSPA Dual antenna equalizer with dominant interferer interference suppression via combination of linear spatial processing and SIC.

2013 LTE FeICIC LTE Variety of linear, maximum-likelihood (ML) quasi-ML, or codeword-iterative techniques applicable.

2015+ LTE NAIC LTE Interference suppression further extended via system side information – e.g. on neighbor cell configuration.

Examples of Standards-Defined Interference Cancelling Advanced Receivers

Device Implications – relentlessly competitive evolution to more advanced predominantly blind interference cancellation

PIMRC 2014 15

Spectral Efficiency Enhancements Opportunities for Innovation

Very Large MIMO (Massive MIMO)

New Modulationand Coding

Non-Orthogonal Multiple Access

Converged Baseband and RF

Very Large MIMO (a.k.a. Massive MIMO) and Joint MIMOOpportunity - exploits very large no. of antennas at (optional) multiple base station to resolve large number of separable spatial sub-channels.Challenge – Channel estimation under low-SNR conditions and maintenance of channel reciprocity; low latency inter-base station coordination.

New Modulation and Coding SchemesOpportunity - Improved spectrum control, reduced L1 time-frequency overhead – e.g. rateless coding, Generalized Frequency Division Multiplexing (GFDM).Challenge – Efficient baseband implementation, efficiency gains.

Non-Orthogonal Multiple Access (NOMA)Opportunity - relaxes time-frequency OFDM orthogonality to increase joint scheduling flexibility, and enhance power amplifier peak-average ratio.Challenge - potentially limited gains over existing modulation and multiple access techniques; requires interference-suppression receivers.

Converged Baseband and RF Architectures and Full-Duplex SystemsOpportunity – Suppression of duplexer leakage, inter-carrier leakage and spurious, inter-RAT interference, transmitter non-linearities to yield flexible wideband transceivers. Combine with co-channel interference suppression to yield full-duplex in-band operation.Challenge – Computation complexity, dynamic range.

mm-Wavemm-WaveOpportunity – Ultra-wideband, ultra-high rate narrow-beam transmission from 28-90GHz.Challenge – Cost, power consumption, stable global spectrum availability.

D2D

D2D (Device-Device)Opportunity - Direct device-device communication including inter-vehicular modes and coordinated interference suppression.Challenge - Traffic amenability to transport mode.

PIMRC 2014 16

mm-Wave System Design Historical Perspective

J. Gardiner, “Microwave and mm-WAVE Technology Requirements…European 4th-Framework…”,

Microwave Symp. Digest, 1995

1995

J.J. O'Reilly, et al, “MODAL: an Enabling Technology for Wireless Access”, 1993. 4th-IEE Conf. Telecomm.

1993

PIMRC 2014 17

Millimeter-Wave Radio Integration Antenna on Substrate

Millimeter Wave Substrates

Antenna Topologies and Architectures

Nor

mal

ized

Rad

iate

d P

ower

[dB

]

Measured Radiation Pattern of LCP Module at 59.5GHz

Example: Phased Array on Liquid Crystal Polymer (LCP)

6 x3 array + Omni antenna Die on Package Back Side Re

ceiv

ed P

ower

[dBm

]

Measurements Simulations

Evaluation of Polarization Diversity

PIMRC 2014 18

Core Mobile Technology Evolution – Potential Future

5

10

15

20

25

2012 2014 2016 2018 2020

Estimated AP Node vs. Year (PQR)

AP Geometry (nm)

AP

Geo

met

ry (n

m)

Year

22nm

14nm 10nm

7nm

Source: Fudzilla

Q3 Q4Q2Q12013 2014 2015 2016 2017 2018 2019

Q3 Q4Q2Q1 Q3 Q4Q2Q1 Q3 Q4Q2Q1 Q3 Q4Q2Q1 Q3 Q4Q2Q1 Q3 Q4Q2Q1

Design Cycle #1

Design Cycle #2

Design Cycle #3

Design Cycle #4

7160 726040nm R10R9 28nm

Atom Z24xx

Cat-6Cat-4

32nm

1-Core 32-Bit

Atom Z25xx32nm

2-Core 32-Bit

2020SoFIA LTE4-Core 64-Bit

Atom Z37xx4-Core 64-Bit

22nm

28nm

0

200

400

600

800

1000

1200

2012 2014 2016 2018 2020

LTE Downlink Data Rate vs. Year

Downlink Data Rate (Mbps)

Dow

nlin

k D

ata

Rate

(Mbp

s)Year

7160(Cat-4)

7260(Cat-6)

1Gbps

Potential Design Region – Ultra Premium Devices

Example Design Cycle Iteration

Simple Linear Extrapolation

Design, Product and Process predictions are speculative and not to be relied upon. Design cycles and process nodes may be realized more or less quickly than noted.

PIMRC 2014 19

LTE Band Support Evolution Looking Ahead at Potential Futures

LTE Bandwidth Support • 7.5bps/Hz efficiency @ 2x2 • 1Gbps requires ~130MHz @ 2x2

~70MHz @ 4x4

0

20

40

60

80

100

2008 2010 2012 2014 2016 2018 2020

Band or Band Combination Support vs. Year

Ban

d or

Ban

d C

ombi

natio

n C

ount

Year

Defined Bands and BandCombinations

Defined Bands Supported Bands

Supported Bands andBand Combinations

LTE Band Support • Growth in defined bands may be slowing • Band combinations growing rapidly • Supporting the emerging set of aggregated

band combinations becoming complex

0

40

80

120

160

200

2012 2014 2016 2018 2020

Aggregated Bandwidth vs. Year

Aggregated Bandwidth (MHz)

Agg

rega

ted

Ban

dwid

th (M

Hz)

Year

20MHz

150MHz

100MHz

40MHz

Simple Linear Extrapolation1

Design, Product and Process predictions are speculative and not to be relied upon. Design cycles and process nodes may be realized more or less quickly than noted.

PIMRC 2014 20

ITU and 5G Requirements ITU-R M.[IMT.VISION]

Attribute IMT-Adv. 4G

IMT-Future 5G

Achievable Rates (Mbps) 1Gbps 10-50Gbps

Connection Density 106-107/km2

Mobility & Coverage 350km/h 500km/h

Energy Efficiency 1x 50-100x

Spectral Efficiency 1x 5-15x

Latency 10ms 1ms

Selected 5G Requirements (ITU WP5D – July 2014)

PIMRC 2014 21

4G Systems Remain Significant in a 5G World Low Cost Markets and Legacy Infrastructure

[1] “Self-Organizing Networks (SON): Self-Planning, Self-Optimization and Self-Healing for GSM, UMTS and LTE”, Hamied et. Al, 2012; [2] Strategy Analytics; [3] Ericsson and Cisco

0

2x103

4x103

6x103

8x103

1x104

1.2x104

2010 2015 2020 2025 2030 2035 2040

Mobile Subscribers by Air InterfaceProjection to 2040

GSMGSM - ProjectedWCDMA/HSPAWCDMA/HSPA - ProjectedLTELTE - ProjectedCDMACDMA - ProjectedXLTE - Projected

Subs

crip

tions

(Mill

ions

)

Year

LTE

WCDMA

GSM

CDMA

5GLTE Continues to Evolve

Speculative

PIMRC 2014 22

RF Proc

FEMWCDMA Rel-15

LTE Rel-15

WiFi – 802.11ax

WiGig – 802.11ad+

GSM/EDGE

“5G”

BT 5.x

GNSS

Location Core

CommsCore

Media Cores

PHY Processing

Sensors

ApplicationCores

Auto Interference Suppression (AIS)Suppress inter-CA or GNSS harmonic and irter-RAT self-interference

Multiple RATs (Radio Access Technologies)Evolution of LTE, HSPA, WiFi, BT basebands, addition of 5G RAT(s)

Multiple GNSS Evolution - GPS, Glonass, Galileo, Beidou, IRNSS, Ancillary Terrestrial Systems

Multi-Band Support>40 LTE bands, >10 HSPA bands, plus WiFi, BT, GNSS bands- Flexible FEM Engineering- 5G mm-wave (10-100GHz) Support

Multi-Antenna Operation- 4-Port+ Operation- Multi-Band, Multi-RAT Port Sharing- Active Impedance Matching

Inter-RAT Coexistence- WiFi, Cellular- Cellular, GNSS

Advanced Baseband Signal Processing- Multiple MIMO modes (e.g. TM10)- MU-MIMO co-scheduled interference suppression- Blind co-channel suppression

Connection Management- WiFi Offloading- Multi-RAT Aggregation

PA Efficiency- Envelope Tracking- Digital Predistortion

Integration Technologies- Flexible Integration and Manufacturing- Flexible Functional Integration

Very Low Power Operation- Advanced Power Management- Delegated Cores

Location Processing- A-GNSS Computation- Sensor Fusion

Advanced Sensors- MEMS

Integration

4G-5GTransitionImpact

High

Medium

Low

IntegrationRFBaseband

Low FrequencyRF (<6GHz)

FEM

mm-WaveDSP

Device of 2020?

Design, Product and Process predictions are speculative and not to be relied upon. Design cycles and process nodes may be realized more or less quickly than noted.

PIMRC 2014 23 PIMRC 2014

Conclusions Key Trends and Pure Conjecture Core Processor Evolution − Driven by continuation of Moore’s Law

Transceiver RF-Baseband Design Innovation − Flexible duplexing; increasing digital signal processing support of RF – potential for full duplexing

Evolution of LTE and WiFi − New performance levels and some extension into new use cases and spectrum

Internet of Things − Drives further growth of WiFi, Bluetooth and cellular IOT access − New low latency, low-complexity wide-area 5G system modalities

Evolved Micro-wave Access − The wide-area ‘core’ of 5G, optimized for low latency and control place stability

mm-wave Access − An integral part of 5G, provided cost, power consumption challenges resolvable

Network Virtualization from Core to Radio Access Network − Enabled by server technology and integrated silicon photonics

These separate elements combine to form the de-facto ‘5G’ To serve a critical new use case…that has not yet been invented.

Thank You

PIMRC 2014 25

PIMRC 2014 26

Moving Towards 5G Technology Timeline

Q3 Q4Q2Q12013 2014 2015 2016 2017 2018 2019 2020

Q3 Q4Q2Q1 Q3 Q4Q2Q1 Q3 Q4Q2Q1 Q3 Q4Q2Q1 Q3 Q4Q2Q1 Q3 Q4Q2Q1 Q3 Q4Q2Q1

ASN.1Dec.‘14

ASN.1Mar.‘16

ASN.11

Dec.‘173GPP

802.11ax TG StartMay ‘14

802.11ax

ASN.11

Sep.‘19

2013 2014 2015 2016 2017 2018 2019 2020

Draft 1.0Jul. ‘16

Draft 2.0Mar. ‘17

802.11ax Amendment1H-2019HEW2

Note 1: Strictly informal estimate of possible date.Note 2: High Efficiency WiFi.Note 3: Potential study and technical working group.

802.11ad+3 802.11ad+ Amendment1

WRC’15Nov. ‘15

WRC’182018

IMT-20202020

Rel-12 Rel-13 Rel-14 Rel-15 Rel-16

IMT-2020

Study1

IMT-2020? IMT-2020?

IMT-2020?

IMT-2020?

PIMRC 2014 27

Internet of Things (IOT) and Wearables Future Growth

PIMRC 2014 28

5G Deployment and Use Cases Creating User Demand

The devices to drive 5G demand, capitalize 5G network deployments and fund 5G device innovation have not yet emerged.

Control at Distance

ImmersiveMultimedia

Transportation Systems

MMI Enabled Communications

Communications2G: Mobile Voice3G/4G: Mobile Data5G: Mobile ?

Control at DistanceLow Latency O(ms)High Speed O(Gbps)

TransportationNavigation & Infotainment: O(Mbps)Autonomous Operation: O(Gbps)

Immersive MultimediaVirtual PresenceEmbedded Multi-view

PIMRC 2014 29

LTE Evolution Candidate Rel-13 Technologies 3GPP Release-13 Potential Technologies

Example Technologies – Release-13

Key Trends Comment

Unlicensed Spectrum Operation

Various scenarios possible including aggregation of licensed and unlicensed spectrum – e.g.: 1. Licensed FDD pair and unlicensed supplemental downlink (SDL) carrier 2. Licensed TDD carrier and unlicensed supplemental TDD carrier

Lightly Licensed Spectrum Operation Arbitrated shared spectrum access – e.g. Licensed shared access (LSA) or 3.5GHz U.S. Spectrum Access System (SAS)

Machine Type Communications (MTC) Examples: 1. Coverage enhancement (+20dB link budget) 2. Cost, complexity and power consumption reduction

Performance Enhancements

Examples: 1. Enhanced throughput (cell edge) – 4-antenna operation at UE 2. Interference rejection (co-channel) 3. Enhanced base station multi-antenna operation (e.g. 3D beamforming) 4. Optimization for specific deployments – e.g. enterprise deployments 5. Cell identification and mobility (possible introduction of ‘phantom’ cell)

WiFi Interworking Examples: 1. Enhanced LTE-WiFi dual access – co-located small cell or network-based 2. Evolved mobility and control procedures

D2D Continuation of Rel-12 D2D work in additional scenarios

PIMRC 2014 30

Unlicensed Operation – WiFi & WiGig Frequency Allocations North America Example

Band Channels Channel

BW (MHz)

Total BW (MHz)

2.4 3 20 60

5.0 21 20 420

60.0 3 2160 6480

Channel BW vs. Band

Allocable Channel Bandwidths Total BW < 6GHz = 480MHz Total BW > 6GHz = 6480MHz Total BW = 6960MHz

PIMRC 2014 31

mm-Wave Frequency Allocations International 60GHz plus U.S. LMDS Bands

Band (GHz)

BW (GHz) Licensing

71-76 5.0 1. Licensing: database registration, non exclusive. 2. Allocation: 1.25GHz aggregable blocks 3. Services: point-point fixed services.

81-86 5.0

92-94 2.0

94.1-95 0.9

U.S. Allocations 70-80-90GHz Bands

BTA1 Lic- ense

Band (GHz)

BW (MHz)

BW (GHz)

A

27.5–28.35 850

1.150 29.1–29.25 150

31.075–31.225 150

B 31.0–31.075 75

150 31.225–31.3 75

Note 1: BTA – Basic Trading Area

U.S. Allocations 28-31GHz LMDS Fixed Point-Point Service

Band (GHz) Structure

38.6-40 14 x 50MHz Pairs (100MHz Total)

U.S. Allocations 39GHz Fixed Point-Point Service

Band (GHz)

BW (GHz) Jurisdiction Licensing

57-64 7.0 USA Unlicensed

57-64 7.0 Canada Unlicensed

59-66 7.0 Japan Unlicensed

57-64 7.0 S. Korea Unlicensed

57-66 9.0 Europe Unlicensed

International 60GHz Allocations

PIMRC 2014 32

Dense Enterprise Deployment Enterprise Coordinated Radio Access Network (C-RAN)

Remote Radio Heads

Remote Radio Heads

Metro-Ethernet Backhaul to Internet (S1 Interface)

Metro-Ethernet Backhaul to Internet (S1 Interface)

Server Coordinated Logical Cells

Server Coordinated Logical Cells

Single Mode Fiber+ Power Distribution

(~1300nm)

Single Mode Fiber+ Power Distribution

(~1300nm)

Routing and Power Distribution

Routing and Power Distribution

PIMRC 2014 33