review article joint channel assignment and routing in multiradio...

Review ArticleJoint Channel Assignment and Routing inMultiradio Multichannel Wireless Mesh NetworksDesign Considerations and Approaches

Omar M Zakaria12 Aisha-Hassan A Hashim1 Wan H Hassan2 Othman O Khalifa1

M Azram1 Lalitha B Jivanadham2 Mistura L Sanni1 and Mahdi Zareei2

1Department of Electrical and Computer Engineering International Islamic University Malaysia 50728 Kuala Lumpur Malaysia2Malaysia-Japan International Institute of Technology (MJIIT) Department of Electronic Systems EngineeringUniversiti Teknologi Malaysia 54100 Kuala Lumpur Jalan Semarak Malaysia

Correspondence should be addressed to Omar M Zakaria dromarzakariaieeeorg

Received 27 November 2015 Accepted 9 May 2016

Academic Editor Sabrina Gaito

Copyright copy 2016 Omar M Zakaria et al This is an open access article distributed under the Creative Commons AttributionLicense which permits unrestricted use distribution and reproduction in any medium provided the original work is properlycited

Multiradio wireless mesh network is a promising architecture that improves the network capacity by exploiting multiple radiochannels concurrently Channel assignment and routing are underlying challenges inmultiradio architectures since both determinethe traffic distribution over links and channels The interdependency between channel assignments and routing promotes towardthe joint solutions for efficient configurationsThis paper presents an in-depth review of the joint approaches of channel assignmentand routing in multiradio wireless mesh networks First the key design issues modeling and approaches are identified anddiscussed Second existing algorithms for joint channel assignment and routing are presented and classified based on the channelassignment types Furthermore the set of reconfiguration algorithms to adapt the network traffic dynamics is also discussed Finallythe paper presents some multiradio practical implementations and test-beds and points out the future research directions

1 Introduction

Wireless Mesh Network (WMN) is a multihop wirelessnetwork characterized by low deployment cost that consistsof mesh routers and Internet gateways It is a special adhoc network with static topologies where unlike the generalad hoc networks mesh routers have no energy limitationsWMN deployments are found in surveillance [1 2] buildingautomation remote healthcare delivery [3 4] and smart grids[5ndash7]Mesh routers are the fundamental part of wirelessmeshnetwork which gives routing support for network trafficof mesh clients Mesh routers can also be equipped withmultiple radios to increase the network capacity and to reducethe interference level over the network The availability andflexibility of IEEE 80211 components make it a good can-didate for wireless mesh deployment 80211-based networksprovide a cheap and flexible wireless access capability andare easy to deploy in campuses airports and hospitals

Backhaul links in 80211-basedWMN can be operated on oneof the several nonoverlapping channels (ie 12 channels for80211a and 3 channels for 80211b) Furthermore the cost-effectiveness of network interface cards made it possible touse multiple radios and channels to increase the throughputMultiple radio configurations allow the utilization of theavailable multiple orthogonal channels (partially overlappedchannel also can be considered as in [8 9]) The networkarchitecture of multiradio wireless mesh networks (MR-WMNs) is illustrated in Figure 1 Multiradio mesh routersare connected through wireless backhaul links over multipleorthogonal channels Mesh clients are connected to meshrouters through different set of links referred to as networkaccess links Mesh clients are user entities with no routingfunctionality In the rest of this paper the term link will referto backhaul link The gateway is a mesh router that connectsthe mesh components with external networks

Hindawi Publishing CorporationJournal of Computer Networks and CommunicationsVolume 2016 Article ID 2769685 24 pageshttpdxdoiorg10115520162769685

2 Journal of Computer Networks and Communications

Multiradio mesh router

Mesh clientBackhaul wireless linkAccess wireless link

Wired link

Internet

CH3

CH3

CH3

CH3

CH3

CH6

CH6

CH6

CH6

CH6CH8

CH8

CH8CH8

CH8CH8

CH8

CH8

CH8

Gateway

Gateway

Figure 1 The multiradio wireless mesh networks architecture

Recently MR-WMN has attracted numerous numbersof research efforts to utilize the advantages that this net-work is offering Several proposed approaches on channelassignment (CA) algorithms multichannel MAC protocolsmultichannel routing metrics links scheduling (LS) mul-tichannel multicast protocols power and topology controland network planning exist in the literature However thedesigns that are considering combinations of these issueshave shown more efficient performance such as joint routingand link scheduling [10] joint CA and power control [11]joint QoS multicast routing and CA [12] joint gatewayselection transmission slot assignment routing and powercontrol [13] joint CA power control and routing [14] jointCA power control and rate assignment [15] joint routing andtopology control with directional antennas [16] and partiallyoverlapped CA [8 9]

CA algorithms aim to assign channels to the radiointerfaces and links with the objective of minimizing theoverall contention and interference over the wireless linksCA can either be solved as a separated problem as in [17 18] orjointly solved with routing as in [19 20] Furthermore it canbe developed as centralized solutions [19 20] and distributedsolutions [21 22] The availability of the entire network viewmakes centralized solution more effective than distributedsolutions which only relayed on local information TypicallyCA algorithms must have knowledge on network load inorder to assign channels to links However some otheralgorithms do not require any traffic information as in [23]where the interference is minimized by conserving a 119896-connected topology

Joint design approaches in WMNs are surveyed in manyworks in the literature such as [24ndash27] However none ofthe works in the literature have comprehensively investigatedthe Centralized Joint Channel Assignment and Routing (C-JCAR) approaches This motivates us to undertake an in-depth review of different C-JCAR proposals in the litera-ture Figure 2 shows the research direction of this paperThe rest of the paper is organized as follows Section 2discusses key design issues models and approaches for C-JCAR algorithms Section 3 presents reviewed works undertwo subsections of C-JCAR approaches and reconfigurationCArouting algorithms Section 4 investigates some prac-tical implantations and test-beds for MR-WMNs FinallySection 5 describes the future research directions and con-cludes the paper

2 Design Issues Modeling andApproaches for C-JCAR

This section identifies the key design issues from link and net-work layers for C-JCAR approaches to address and clarifiesinterferencemodel and networkmodel to be used in problemformulation and algorithms It also presents themathematicalformulation of the problem including optimization objec-tives constraints and fairness Besides this section also givesthe general classifications to C-JCAR algorithms

21 Key Design Issues211 Channel Assignment Schemes Based on the underlinghardware CA can be implemented in one of three schemes

Journal of Computer Networks and Communications 3

Organization of the paper

1

2

3

Design issues modeling andapproaches for C-JCAR

Related works

Summary practicalimplementation platformsfuture directions

Design issues

Modeling

Approaches

CA typesTraffic type and routing schemesConnectivity and physical topology

Interference modelingNetwork modeling

Optimization functions and fairness Mathematical optimization constraintsAlgorithms

Static CAStatic CA with link schedulingDynamic CA

Reconfiguration

C-JCAR

Channel reassignmentReroutingReconfiguration trigger

Summary of related works (Tables 2 and 3)Practical implementations and test-bedsConclusion and future research directions

Figure 2 Organization of the paper

If the implemented technology does not support any link-level synchronization then only Static CA (SCA) can beimplemented and channels are assigned to radio interfacesfor a long term However if synchronous coordination issupported at link layer then SCA can be implemented withlink scheduling (LS) and the wireless link can be active onspecific time slots only let us refer to this case by SCA-LS Dynamic CA (DCA) is the general case where radiointerfaces are capable of switching their channel in a smalltime compared with the time slots Thus wireless linkscan operate on different channels at different time slotsTo illustrate the difference between the three CA schemesFigure 3 shows the possible channel allocation on a wirelesslink As it is shown in Figure 3 for SCA links are allowed tobe active on one channel only at all times In SCA-LS linksare allowed to be active on one channel only and on specifictime slots However if DCA is adopted links can operate ondifferent channels on different time slots

212 Traffic and Routing Issues As in load-aware algorithmsthe aggregated traffic loads over mesh routers are requiredas an input This traffic information can either be mea-suredestimated online as in [28 29] or assumed based on ahistorical profile [19]WMNs can be used to deliver two typesof traffic the Internet traffic and peer-to-peer traffic Internettraffic is the traffic received or directed to the Internet [2030ndash33] Either outbound traffic (to the gateway) [20 30] orinbound traffic (from the gateway) [31] can be considered infinding the network configuration In [28] specific software

such as IPFIX system [34] is used at the gateway to collect thetraffic information The upper and lower bound of the trafficdemand are assumed to be available for each mesh router[33] A traffic profiler at each mesh router collects the trafficinformation and sends it to a centralized entity [29 35] wherepeer-to-peer traffic is assumed References [36 37] assumethe traffic load is elastic and only information on source-destination communication pairs are considered

Routing algorithms in wireless ad hoc networks are cat-egorized as proactive (table-driven) reactive (on-demands)and hybrid algorithms In proactive algorithms such as [38ndash40] routing table is built individually through exchangingrouting information with other nodes in the network Inreactive algorithms such as [41ndash43] no routing tables aremaintained and instead each node triggers the route dis-covery process whenever it has traffic to deliver to a desti-nation Hybrid algorithms [44ndash47] implement the conceptsof proactive and reactive routing protocols Ad hoc routingprotocols are extended to be used in WMNs The routingdecisions in these protocols are decentralized process andeach node is responsible to make its routing decision On theother hand in centralized routing paradigm routing decisionis performed at a centralized entity and nodes build theirroutingforwarding table based on the updates received froma centralized entity In WMNs if traffic is identified permesh client then mesh routers must maintain a forwardingentry for each mesh client and layer-2 addresses are usedto identify each flow This is similar to 80211s standardwith Hybrid Wireless Mesh Protocol (HWMP) where path


SCA

SCA-LS

DCA Time(slotted)

Time(slotted)

Time

Ch1

Ch1 Ch1Ch8 Ch8

Ch1

Ch1

Ch1

Ch5

Figure 3 CA types on a wireless link

selection is based on layer two addressing However pathselection is usually associated with scalability issues On theother hand if the traffic is identified per mesh router andflow is defined as aggregated traffic that originated from onemesh router to another mesh router then layer-3 routing isused Inmost cases centralized routing approaches follow thelatter definition of flow A TCP-level flow path selection isassumed in [28] However this introduces more overheadsto the routing layer Centralized routing proposals builtrouting tables based on source-destination manner whereboth source and destination addresses are required in therouting decision With centralized routing in MR-WMNmodification is required on the routing table structure andthe Address Resolution Protocol (ARP) However a compli-cated routing scheme is required for dynamic CA especiallyif no static binding is applied between radio interfaces andneighbors For better exploitation of the topology structureof MR-WMN multipath routing is assumed in most relatedworks to achieve load balancing However the out-of-orderproblem and the configuration complexity are the maindrawbacks of multipath scheme

213 Topology and Connectivity Issues Several mechanismshave been used in the literature to control the topologyformations in WMNs This includes power control [48ndash50]the use of directional antennas [16 51] and routing [23 52ndash54] Furthermore in multiradio architecture CA is anotherfactor that determines the physically connected topology Anoverview of the topology control mechanisms and issues inMR-WMNs is presented in [55] In centralized algorithmsthe full network topology is assumed to be available at thecentralized entity Topology can be obtained using existingrouting protocol such as OLSR [28] or the network manage-ment protocol as in [32] In order to define different possibletypes of topology formation let us define logical links as theset of potential links that directly connect mesh routers ifproper channels are assigned to their radios and let us definethe physical links as the actual wireless linkswith a designatedchannels assigned to it Since inappropriatemapping can leadto disjoint topology mapping the logical link onto physicallink must be carefully determined Furthermore assigningchannels to all logical links affects the channel diversity onthe networkThis is due to node-radio constraint (elaboratedmore in next subsections) more specifically with a smallnumber of radio interfaces On the other hand physical linksneed to be mapped onto the active links where active linksare the set of links carrying traffic and are determined by

routing algorithm Figure 4 shows two types of physical andlogical links mapping In Figure 4(a) channels are assignedto each logical link as in [19 29 37] In Figure 4(b) channelsare assigned to a number of logical links only as in [2036 56 57] However some works allow multiple physicallinks to exist between adjacent mesh routers [30 56] whichneed to be considered in the routing procedure Meanwhileconnectivity must also be addressed in CA algorithms toprevent isolating parts of mesh routers from receiving thecontrol messages and the configuration updates from thecentralized controller Connectivity is achieved in [19 29 37]by assigning a channel to each virtual link or by dedicating aradio on a common channel at each mesh router [58]

22 Interference and Network Modeling221 Interference Modeling Interference plays an importantrole in wireless networks and has a significant impacton the network performance MR-WMNs are proposed toreduce the contention on the communication channel andto distribute the transmission over several channels CAalgorithms tend to exploit the channel diversity to reducethe interference levels in the network Several interferencemodels are proposed in the literature to model the interfacein the wireless networks Three different interference modelsare presented in this section namely the interference-rangemodel the protocol model of interference and the physicalmodel of interference In order to explore these interferencemodels let us assume that all radios in the network have fixedtransmission power with omnidirectional antennas and thesignal propagation model is based only on frequency anddistance Each radiowill have a fixed transmission-range (119877

119879)

and two radios can form a wireless link if they are withinthe transmission-range of each other see Figure 5(a) Thereceiver radio at each wireless link can correctly receive anddecode the transmission in the absence of any interferingradio

The interference-range model is the simplest modelwhere each radio has its interference range (119877

119868) where 119877

119868=

Δ sdot 119877119879 Δ gt 1 (default Δ = 2) A directional transmission

on a wireless link is successfully completed if the receivingradio is not within the interference range of another activetransmitting radio as shown in Figure 5(b) Thus in order tosuccessfully receive the transmitted signal from radio (V) toradio (119909) on wireless link (V 119909) (1) must be true

1003816100381610038161003816119895 minus 1199091003816100381610038161003816 ge 119877119868 forall119895 isin Radio Interfaces V

|V minus 119909| le 119877119879(1)


Physical wireless link

Multiradio wireless mesh router

(a) 1-to-1 mapping

Logical wireless link



(b) 1-to-any mapping

Figure 4 Topology formation logical-to-physical links mapping

Transmission range

Wireless link

z

n

m

y

u

x

(a)

Interference range

Interferingm

n

z

y

u

x

transmission

(b)

Figure 5 (a) Transmission-range and (b) interference-range model

On the other hand interference in protocol model doesnot assume a fixed interference range Instead the interfer-ence is determined based on the Euclidean distance betweeninterference radios transmitting radio and receiver radioThus a directional transmission on a wireless link can besuccessfully completed if the distance between the interferingradio and the receiving radio is larger than the linkrsquos lengthThus in order to successfully receive a signal transmittedfrom radio V to radio 119909 on wireless link (V 119909) in protocolmodel (see Figure 6(a)) (2) must be true

1003816100381610038161003816119895 minus 1199091003816100381610038161003816 ge Δ sdot |V minus 119909|

forall119895 isin Radio Interfaces V Δ gt 1

|V minus 119909| le 119877119879

(2)

In comparisonwith the previous twomodels the physicalmodel is the most realistic model where concurrent trans-missions from multiple interfering radios are accounted in

this model In order to successfully receive the transmittedsignal the SINR (signal-to-interference-and-noise ratio) atthe receiving radio must be greater than a predefined thresh-old Thus V radio can successfully receive the transmittedsignal from radio 119909 on wireless link (V 119909) (see Figure 6(b))if (3) is true

119875V

119873 + sum119895isinradiosV 119875119895

ge 120573V119909 (3)

where119873 is the thermal noise power in the frequency channel120573V119909 is an SINR-threshold which is determined based on theconsidered modulation and coding schemes and 119875V is thereceived power from V at 119909 SINR-threshold is chosen in away that the resulting BER is higher than acceptable BERby the modulating technique SINR-threshold is based onthe experimental observation and can be mapped onto a biterror probability [59] The physical model of interference isextended in [60] to include the shadowing effect of the RFsignals The work of [60] studied the minimum required


Interfering transmission

n

m

z

y

u

x

| minus x||nminusx|

|uminusx|

|n minus x| gt Δ middot | minus x|

|u minus x| lt Δ middot | minus x|

(a)

Received signalstrengths from

Transmissions contribute to theinterference at the receiver

Received signal from

n

m

y

z

u

x

u y z n m

(b)

Figure 6 Interference based on (a) protocol model and (b) physical model


Interference range

z

n

m

y

u

x

Figure 7 Interference based on interference-rangemodel for 80211

number of channels to achieve interference-free channelassignments under realistic interference model called SINRmodel with shadow

In the conventional 80211 MAC protocol with CarrierSense Multiple Access-Collision Avoidance (CSMACA)wireless links are bidirectional since the sender still needsto receive acknowledgement messages from the receiver (ifRTS-CTS is enabled it also receives CTSmessage)Thereforeall transmissions causes that interfere with the sender or thereceiver must be avoided for successful transmission In thismodel sending and receiving nodes on a link are potentialsource of interference to another link transmission Figure 7shows the set interfering radios to wireless link (V 119909) basedon the 80211 bidirectional model

222 NetworkModeling Network inMR-WMNarchitecturecan bemodeled using either a router-to-routermodel [20 5657] or a radio-to-radio model [31 36 61] In router-to-routermodel the network is modeled as a directed graph 119866(119881 119864)where 119881 represents the set of mesh routers in the networksand 119864 is the set of directed links A directional link existsbetween twomesh routers if direct communication is possiblebetween them In radio-to-radio model vertices representthe radio interfaces A link exists between two verticesif radios can communicate directly and are placed in twoadjacent mesh routers Radio-to-radio model is usually usedif the radio interfaces are heterogeneous and support differenttechnologies or if different radios operate on different subsetsof channels Figure 8 illustrates the difference between radio-to-radio and router-to-router model for single channel Formultichannel scenarios each link between two radiosmust bereplicated into the number of available orthogonal channelssupported by the mesh router or radio pairs [31]

23 Problem Formulation and Approaches for C-JCAR

231 Optimization Objective Functions and Fairness The C-JCAR algorithms have been designed in the literature fordifferent optimization objectives andmetrics below is the setof optimization objective functions and metrics

(i) Maximize aggregated network throughput [29 31 3336 62]

(ii) Maximize achievable scaling factor (120582) [20 30 31 5661]

(iii) Maximize the aggregated utility of flows [37]

(iv) Minimize the maximum interference on all channels[20 63]


Router-to-router model

Radio-to-radio model

Directional link


Radio interface

Figure 8 Router-to-router versus radio-to-radio models

(v) Maximize the minimum unutilized capacity on links[57]

(vi) Maximize the cross-section good-put [19](vii) Minimize the path length and link contention [32

64]

Fairness is also another objective to be considered Fair-ness ensures a fair resource allocation among network usersor traffic Several fairness constraints are introduced for C-JCAR To elaborate the fairness constraints let us assume FL isthe set of flows in the network and let120596

119891 0119891be the demanded

load and the actual achieved load of flow 119891 isin FL the fourtypes of fairness constraints in the literature are as follows

(i) Min-120582 Fairness This constraint is to ensure that eachflow achieves at least 120582 of its demands The Min-120582 fairnessconstraint is presented in

0119891ge 120582 lowast 120596

119891forall119891 isin FL (4)

(ii) Max-to-Min Fairness This constraint is to constrain thedeference between the highest and the lowest achieved loadsThen for a given 120583 the Max-to-Min fairness constraint isgiven in

120583 ge0max0min

(5)

where 0min is the lowest and 0max is the highest load

(iii) Proportional Fairness This fairness can be used as anoptimization objective function and it is given as in

maximize sum

119891isinFLlog (0

119891) (6)

(iv) Bounded Fairness This fairness configuration boundseach flow with specific upper and lower bound

Min-120582 fairness is considered in [20 29ndash31 56 61] whereflows may have different demands Other fairness constraintssuch as Max-to-Min fairness proportional fairness andbounded fairness are used in [37 62] and [33] respectively

232 Mathematical Formulation Constraints Thegeneral setof constraints to be considered in the mathematical formula-tion of the C-JCAR problem can be divided into six sets asfollows

(1) Flow Routing Conservation Constraints These constraintsensure that for each traffic flow 119891 from source to destinationthe net amount of traffic out of each mesh router is equalto the flow rate (119891

119903) if the mesh router is the source of the

traffic and (minus119891119903) if it is the destination and otherwise 0 More

constraints can be added to the problem to determine therouting scheme (for multipath or single-path)

(2) Radio Constraints These constraints ensure that thenumber of channels assigned to each incident link on a meshrouter (at any given time slot for dynamic CA) does notexceed the number of radio interfaces at that node Thisinteger-value constraint can be relaxed to linear constraintsto ensure that the total load on the links on each node is nothigher than the number of radios on that node multiplied bythe channel capacity

(3) Interference Constraints (or Capacity ConstraintsSched-ule-Ability Constraints) These constraints ensure that theamount of data flows on interfered links does not exceed aspecific value This is to constraint the maximum contentionlevel over the collision domains if contention-based MACis considered or to ensure a feasible interference-free linkscheduling if contention-free MAC is used

(4) Link Capacity Constraints These constraints ensure thatthe traffic load on a wireless link is not exceeding thelink capacity This constraint is implicitly considered underthe interference constraints if only sufficient condition forschedule-ability is considered

(5) Fairness Constraints These constraints ensure fairness inallocating resources to different traffic flow demands

Apart from the above-mentioned constrained otherconstraints on the topology can also be introduced Theremaining part of this section will be discussed with regard tothe radio and interference constraints The Notations depictsa set of notations for mathematical formulation employed inthe derivation


If CA and interference-free scheduling is computed forall links in time slotted system the allocated capacity of awireless link 119897 can be obtained by

Link capacity (119897) =119905119897

119879sdot 119862119897 (7)

where 119905119897is number of time slots when 119897 is active 119879 is the

total number of time slots and 119862119897is the channel capacity at

link 119897 However in contention-based networks and withoutsupported time slotting it is not possible to allocate adedicated bandwidth (capacity) to a link Instead virtual linkcapacitymay be estimated for each active link By consideringthat the link capacity is equal for all links and all links canutilize the maximum channel capacity (119862

119874) with the absence

of interference transmission the virtual capacity for each link119897 can be calculated as

VC (119897) =119892 (119897)

sum1198971015840isin119868(119897)cup119897

119892 (1198971015840)sdot 119862119874 119897 isin 119871 (8)

In [57] the virtual capacity (effective capacity) of a link isdefined as follows

VC (119897) + sum

1198971015840isin119868(119897)

VC (1198971015840) = 119862119874

forall1198971015840isin 119871 (9)

The link capacity constraints in [19 57] are given as follows

119892 (119897)

VC (119897)le 1 forall119897 isin 119871 (10)

To find a virtual link capacity independent set of edges isconsidered in [37] however this approach is not applicablein multiradio multichannel since radio constraints are notconsidered in the channel allocation for the independent sets

Among the interference models mentioned in the previ-ous section the protocol and interference-range models arebinary pairwise models where the interference between anytwo links either exists or does not exist The interferenceconstraints with necessary condition for interference-freescheduling can be presented as

sum

1198971015840isin119868(119897)cup119897

119892 (1198971015840)

1198621198971015840

le 120579 forall119897 isin 119871 (11)

where 120579 is the interference constant that depends only on theinterference model For instance 120579 in the interference-rangemodel is a function of the ratio between the transmission andinterference range (119877

119879119877119868) where it represents themaximum

number of links interfering with a specific link but nota pairwise interfere Interference constraints can also beconsidered for sufficient condition only as follows

sum

1198971015840isin119868(119897)cup119897

119892 (1198971015840)

1198621198971015840

le 1 forall119897 isin 119871 (12)

Thus for any solution that satisfies the sufficient condi-tion there is a feasible link scheduling (proof is stated in[20]) The interference constraint with sufficient condition is

Time slots

InterferenceRadio interface

Wireless link

Link

s e2

e5

e1

e4e3

e2(03)

e5(04)e1(02)

e4(03)

e3(04)

Figure 9 Example of links scheduling (sufficient condition is notsatisfied)

used in [30 33 36 56 57 62] Figure 9 demonstrates that itis possible to have feasible interference-free links schedulingeven though only necessary conditions are satisfied Let usassume that the channel capacity is equal to 1 and the numberthat associated with each link represents the amount of trafficload carried on each link

From Figure 9 the link flows do not satisfy the sufficientcondition of link scheduling (larger than 1) However afeasible scheduling exists as obtained in Figure 9 It can beconcluded from the previous example that solutions basedsufficient conditions will have an optimality gap Using thenotation (see Notations) interference and radio constraintswith different CA schemes are presented in Table 1

Another way to model the interference on wireless meshnetwork is by using empirical interferencemeasurementThiscan be done by taking the advantage of the static natureof WMN where the required information can be easilymeasured and recorded beforehand Empirical interferencemeasurement is presented in [37]Thepacket error rate (PER)of a wireless link at a receiver is empirically determinedwhen another link is active Based on this measurement theobjective of the CA algorithm is to minimize the summationof the PER of one linkwhen another link is activelymultipliedwith the load on each linkThe objective function in [37] is asfollows

Minimize sum

1198971

sum

1198972isin119868(1198971)

119892 (1198971) lowast 119892 (119897

2) lowast PER (119897

1| 1198972) (13)


Table 1 Mathematical formulation (interference amp radio constraints)

Operating mode Network model Interference (capacity) constraints Node-radio constraints

SCA


general

LPNecessary condition

sum

1198971015840isin119868(119897)cup119897

119892 (1198971015840)

1198621198971015840

le 120579 forall119897 isin 119871

Sufficient condition

sum

1198971015840isin119868(119897)

119892 (1198971015840)

1198621198971015840

le 1 forall119897 isin 119871

IPsum

119894isin119862

ℎ119894

119898le 119870 (119906) forall119906 isin 119873

LPsum

119897

ℎ(119897)=119898

or 119905(119897)=119898

119892 (119897)

119862119897

le 119870 (119906) forall119906 isin 119873


[37]

LPLink virtual capacity (effective capacity)sum

1198971015840isin119868(119897)cup119897

VC (1198971015840) le 119862119897 forall119897 isin 119871amp 119892 (119897) gt 0

119892 (119897)

VC (119897)le 1 forall119897 isin 119871

DCASCA-LS


general

IPNecessary amp sufficient condition119910119905(119897) + 119910

119905(1198971015840) le 1 forall119897 isin 119871amp 1198971015840 isin 119868 (119897)

Sufficient conditionsum

1198971015840isin119868(119897)cup119897

119910119905(1198971015840) le 1 forall119897 isin 119871 119905 isin 119879

Or LP relaxationNecessary condition

sum

1198971015840isin119868(119897)cup119897

119892 (1198971015840)

1198621198971015840

le 120579 forall119897 isin 119871


sum

1198971015840isin119868(119897)

119892 (1198971015840)

1198621198971015840


IPsum

119897isin119871

119905(119897)=119906

or ℎ(119897)=119906

119910119905(119897) le 119870(119906) forall119906 isin 119873

LPsum

119897isin119871

119905(119897)=119906

or ℎ(119897)=119906

119892 (119897)

119862119897

le 119870 (119906) forall119906 isin 119873


[56]

IPNecessary amp sufficient conditionIf 119899119898 interfere with each other

sum

1198971015840isin

119864(119899)cup119864(119898)


LPSufficient condition

sum

1198971015840isin119868(119897)

119892 (1198971015840)

1198621198971015840


Radio-to-radiomodel [31]


119905(1198971015840

) le 1 forall1198971015840

isin 119868 (119897)


sum

1015840

isin119868(119897)

119892 (1198971015840

)

1198621015840

le 1 forall119897 isin

IPsum

isin119871

119905(119897)=119906

or ℎ(119897)=119906

119910119905(119897) le 119870 (119906) forall119906 isin 119873

LP

sum

isin

119905(119897)=119903

or ℎ(119897)=119903

119892 (119897)

119862



LPSufficient condition (both interference and radio constraints are considered)

sum

1015840

isinIR(119897)cup119897

119892 (1198971015840

)

1198621015840



Each critical maximum independent set (MIS) is a set of links which satisfyinterference and radio constraints and only one critical independent set is allowedto be active at any time slot Therefore interference and radio constraints areimplicitly considered in the LP formulation


Connectivity

Link mapping

Topology info

Channelassignment

Interferencemodel

Technology

Traffic type

Traffic profile

Traffic collection

Routing scheme

Routing table

Execution (Re-)

Optimizationobjective

Fairnessconsideration

Algorithms

Conserved by assign channel toeach logical link radio oncommon channel

Collected by routing protocolnetwork management protocol

Assign channels to active linksonly to all links 1-to-many

Protocol model physical modelRTS-CTS model interference-range model empiricalmeasurement

Static static with linkscheduling dynamic

CSMA time slotted modenumber of available channels

Multipath single-path multipathmultigateway

Source-based routing source-destination-based routing table25-layer forwarding paradigm

Src-Dest MRs peers load demandsbetween MRs peers sequence ofload for Src-Des MRs peers

Peer-to-peer Internet-based traffic

Traffic monitoringat gateway traffic profiler at eachMR historical profilepredetermined (user-paid pricesupper and lower bound)

Heuristic and iteration-basedmetaheuristic mathematicaloptimization methods andsolvers

Maximize aggregated throughputmaximize achievable scaling factor ofdemands maximize aggregatedutility of flows maximize the cross-section good-put minimize maxinterference and contention

Offline planning stage online state-aware interference levelload distribution interferencelevel periodic bases

fairness proportional fairnessmaximum and minimumbounded constrains

Topology

Approaches Traffic

Routing layer

Link layer

Centralized CAand routingdesign issues

Min-120582 fairness max-to-min

Figure 10 C-JCAR key design issues and approaches

where PER(1198971| 1198972) is the PER at receiver node of link 119897

1when

1198972is activeTo present the interference constraints using the physical

model of interference two approaches can be used as pre-sented in [62] The first approach proposed in [65] is basedon length classes that distributed links to different sets basedon their lengths For each length set (119895) the area is dividedinto a set of square grid cells 119860

119895 Let 119860 isin 119860

119895 Δ(119860) be the set

of links from 119895 with their receiver lying inside 119860 and operateon the same channel The interference constraints are thenpresented as

sum

119890isinΔ(119860)

119909 (119897) le 120593 forall119895 119860 isin 119860119895 (14)

where 120593 is a constant number and 119909(119897) is 1 if the link is activeand 0 if otherwise

The second model is weight-based model where 11988211989721198971

isthe receiving power at 119897

2receiver when both links are active

on the same channel thus

1198821198972

1198971

= 120573 lowast119875V (119903)

119875V (119906)

forall1198971 1198972isin 119871 amp 119897

1= (119903 119904 119888) 1198972 = (119906 V 119888)

(15)

where 119875119909(119910) is the transmission power received at 119909 when 119910

is transmitting the signal The interference constraint can bepresented as

119892 (1198971) + sum

1198972isin1198711198971ampCh(119897

2)=Ch(119897

1)

119873(1198821198972

1198971

) sdot 119892 (1198972) le 120599 (119897

1)

forall1198971isin 119871

(16)

where119873(119909) is the normalized value of 119909 and 120599(119897) is a constantbased on the receiving power at the receiver when 119897 is active

233 Approaches for C-JCAR Thegeneral formulation of theC-JCAR problem can be presented as Integer Linear Pro-gramming (ILP) problem However the CA problem whichis part of C-JCAR problem is an NP-hard [19] where withgiven link loads finding the CA for a set of radio interfacessuch that the link loads are schedulable is NP-hard problemMoreover interference-free link scheduling problem for agiven link flows is also NP-hard [66] Therefore the problemis then transformed into a linear programming problem byrelaxing some constraints [20 31 36 56 62]The problem canthen be considered as multichannel multicommodity flowproblem with constraints on radios interference and linkcapacities The optimal solutions are obtained for feasibleproblems only [33 57] or with some fairness considerations[20 56] Due to the NP-hardness of the problem standardmethods and solvers are not able to solve the problem forreal-life network sizes A set of heuristic algorithms [19 2937 58] and metaheuristic algorithms [33 57] are proposedin the literature [20 31 56] Furthermore approaches canalso be classified based on their execution time into online[29 37] and offline for planning stage [33] Figure 10 showsthe set of design consideration classification and methodsin designing a load-aware C-JCAR proposal

3 Related Works

In this section related works for CA and routing are pre-sented in two subsections The related works in the literatureto solve C-JCAR are presented in the first subsection Theworks are classified based on their CA scheme (SCA SCA-LS and DCA) into three categories This is followed by acomprehensive summary table (Table 2) that compares allapproaches based on the design issues presented in previouspart of the paper and highlights their assumptions and


Table2C-

JCARprop

osalssum

mary

Work

Link

layera

ndchannelassignm

ent

Traffi

camprouting

Topo

logy

Objectiv

eampfairn

ess

Noteslimitatio

nsCA

type

Interfe

rence

Techno

logy

Traffi

ctype

Traffi

cload

Routingscheme

Mapping

(1-to-119909)119909

Con

nectivity

Objectiv

eFairn

ess

[19]

SCA

B-IRM

80211

P2P

Yes

SPM

P1

Yes

Maxim

izec

ross-sectio

ngood

-put

No

Optim

ality

isno

tguaranteedNoaccuratelin

kcapacityestim

ationin

aheavy

traffi

cloadLess

channeld

iversitydu

etolin

kmapping

[56]

DCA

SC

A-LS

B-IRM

mdashP2

PYes

MP

01

Yes

Maxim

ize120582

(scalin

gfactor)

Min-120582

Out-of-o

rder

prob

lem

andconfi

guratio

ncomplexity

onroutingSw

itching

overhead

issues

[20]

SCA-

LSB-IRM

mdashInternet

Yes

MP

Any

ForM

Rswtr

affic

Maxim

ize120582

(scalin

gfactor)

Min-120582

Radioconstraintsa

reno

tcon

sidered

intheL

POut-of-o

rder

prob

lem

duetomultip

athrouting

[36]

DCA

B-IRM

mdashP2

Psrc-desM

RsMP

Any

No

Maxim

izea

ggregated

throug

hput

No

Out-of-o

rder

prob

lem

andconfi

guratio

ncomplexity

onroutingSw

itching

overhead

issues

Optim

ality

issue

since

onlysufficientcon

ditio

nsarec

onsid

ered

[57]

SCA

B-IRM

80211

P2PInternet

Yes

SPAny

ForM

Rswtr

affic

Maxim

izem

insparelink

capacity

Min-120582

onlin

kutilizatio

n

Usin

gsufficientcon

ditio

nin

MILPmay

notlead

toglob

alop

timalsolutio

nSing

lechannelinitia

ltopo

logy

may

lead

toinfeasiblesolutio

nin

heavy

traffi

cloadscenarios

[30]

SCA-

LSTH

MInternet

Yes

MPpredefined

01

ForM

Rswtr

affic

Maxim

ize120582

(scalin

gfactor)

Min-120582

Con

sideringon

lynecessarycond

ition

may

lead

toinfeasiblesolutio

nOut-of-o

rder

prob

lem

dueto

multip

athrouting

[70]

SCA

PHM

(wo

accumulation)

80211

Internet

No

MP(M

-GW)

L25forw

arding

ge1

Yes

Minim

izem

axlin

kutilizatio

nNo

Radioor

interfe

rencec

onstr

aintsa

reno

taccoun

tedin

initialtraffi

cloadestim

ation

Assum

esam

etrafficd

emands

inallm

eshrouters

Out-of-o

rder

prob

lem

duetomultip

athrouting

Link

rateassig

nmentisc

onsid

ered

[31]

DCA

PRMIRM

Internet

Yes

MP(M

-GW)

Any

ForM

Rswtr

affic

Maxim

izea

ggregated

throug

hput

Min-120582

Out-of-o

rder

prob

lem

andconfi

guratio

ncomplexity

onroutingEa

chradiocanbe

operated

onas

ubseto

fchann

elsC

ombinatio

nof

multip

lechannelsto

singlec

hann

elispo

ssible

[32]

SCA

THM

Internet

Yes

SPAny

Yes

Metric

onho

pcoun

tampinterfe

rence

No

Solvingsubp

roblem

with

localcon

straintsdo

esno

tguarantee

glob

alop

timalityInformationon

received

power

atan

odew

henanothern

odeis

transm

ittingisassumed

available

[62]

DCA

IRMP

RM

PHM

P2P

No

SPM

PAny

No

Maxim

izea

ggregated

throug

hput

Max-to

-Min

Equalloaddemandisassumed

atmessrou

tersIn

single-path

routingop

timality

isno

tguaranteed

since

mathematicalform

ulationisform

ultip

ath

[61]

DCA

B-IRM

P2P

Yes

MP

Any

No

Maxim

ize120582

(scalin

gfactor)

Min-120582

Theq

ualityof

thes

olutionishigh

lydepend

ento

nthep

erform

ance

ofcriticalM

ISfin

ding

algorithm

[33]

SCA

B-IRM

80211

Internet

Lower

ampup

per

boun

dMP(M

GW)

Any

ForM

Rswtr

affic

Maxim

izea

ggregated

throug

hput

Boun

dedfairn

ess

Develo

pedforp

lann

ingsta

gewith

different

resourcesc

onstr

aintsNot

practic

alforlarge

networks

since

fitnessfunctio

ninvolves

solving

entiren

etworkLP

[29]

SCA-

LSTH

MP2

PYes

MP(predefin

ed)

1Yes

Maxim

izea

ggrachieved

demands

No

Link

loadsu

sedto

theC

Aaree

stimated

with

out

consideringradioor

interfe

rencec

onstr

aintsLess

channeld

iversitydu

etolin

kmapping


Table2Con

tinued

Work

Link

layera

ndchannelassignm

ent

Traffi

camprouting

Topo

logy

Objectiv

eampfairn

ess

Noteslimitatio

nsCA

type

Interfe

rence

Techno

logy

Traffi

ctype

Traffi

cload

Routingscheme

Mapping

(1-to-119909)119909

Con

nectivity

Objectiv

eFairn

ess

[37]

SCA

EMInternet

Num

bero

factiv

eflow

sMP(SPper

follo

w)

1Yes

Minim

izec

ontention

(CA)

No

Link

loadse

stimated

with

outcon

sideringradioor

interfe

rencec

onstr

aintsNot

clearlysta

tedwhere

andho

wto

implem

entfl

owratecontrol

[63]

SCA

B-IRM

80211

Internet

Yessequ

ence

oftraffi

cMP

Any

ForM

Rswtr

affic

Minim

izethe

maxim

umcongestio

n(ofw

orst

case)

No

Num

bero

ffacetsincreasee

xpon

entia

llyin

large-scalen

etworksthisincreases

then

umbero

fconstraintsinLP

prob

lemhow

everrelaxingthe

convex

hullto

boxrang

esaffectsthee

fficiency

NoteSCA

static

channelassignm

entDCA

dyn

amicchannelassignm

entLS

link

schedu

lingIM

interfe

rencem

odel

IRMin

terfe

rence-rangem

odel

PRMprotocolm

odel

PHMphysic

almod

elB-bidire

ctional

THMtwo-ho

pmod

elEM

empiric

almeasurementP2

Ppeer-to

-peerSP

single-pathM

Pmultip

athMGWm

ultig

atew

ayM

Rmeshrouterand

LLlinklayer


limitations The second subsection presents the set of CAor routing proposal which deals with the reconfigurationproblem (state-aware proposals) These works consider thecurrent network configuration in finding the next one toreduce the overhead associated with the reconfigurationprocess The works are divided into channel reassignmentproposals traffic rerouting proposals and the reconfigu-ration policy The latter one discusses some policies thatdetermine when to trigger the reconfiguration process Thissubsection followed by a summary table (Table 3) discussesstate-aware works and highlights the approaches and thelimitation of each work

31 C-JCAR Proposals

311 C-JCAR Based on SCA In [19] a heuristic algorithmis developed to solve the C-JCAR problem Given the trafficload information for set of source-destination pairs of nodesthe algorithmfinds the routing paths for each flowand assignschannels to virtual links Multipath and single-path routingschemes are considered The optimization objective in thiswork is to maximize the cross-section good-put which isdetermined by matching the expected load on links withvirtual capacity The virtual capacity of a link is estimatedbased on the total load of the interfering linksThe algorithmiterates over CA and routing steps to improve the overallcross-section good-put The algorithm retains the networkconfiguration and tries to improve the configuration byrerouting nonconforming flows When the allocated virtualbandwidthmeets the flow demands for each flow or when themaximum number of iterations is reached the algorithm isterminated

C-JCAR problem in 80211-based wireless mesh networkis formulated as Mixed-Integer Linear Programming (MILP)in [57] In this work the virtual capacities are assigned tological links The MILP minimizes the maximum contentionon the channel assigned to each link by maximizing theminimum difference between the effective capacity and thetraffic load on each link This leads to the reduction inoverall regions interference across the network For small-scale network MILP can be solved using commercial soft-ware such as CPLEX [67] However this is not feasiblefor large-scale networks Therefore metaheuristic algorithmbased on Iterated Local Search (ILS) is proposed to solvethe MILP problem This algorithm commences with logicaltopology consisting of single channel assigned to all logicallinks Then MILP is solved iteratively using relaxed binaryrouting variable to simplify computation of the cost In eachiteration a set of variables is considered as a constant whichtakes their values from previous iteration till one link isselected randomly based on worse case congestion rankingThe next phase involves finding based on the maximumrelaxed routing variable values However the algorithm in[57] starts the solution from a single channel assumptionwith limited network capacity and this can result in aninfeasible solution if heavy traffic load is applied Thereforein [68] Tabu search with iterated local search algorithm isproposed to overcome the above-mentioned problem in [57]The Tabu-search based algorithm selects the solution with

minimum number of conflicting links in each iteration andthe remaining procedure in the algorithm is similar to [57]Furthermore [69] has extended the work of [57] to supportdirectional antenna to reduce the interference in the topology

The CA is jointly solved with link transmission rateassignment in [70] A centralized greedy heuristic is designedto assign channels to given schedulable flow rates Theheuristics algorithm starts off with throughput optimizationfor a precomputed flow rate The initial traffic rate onlinks is computed by assuming interference-free links withobjective to maximum achievable aggregated throughput onthe network and then solves the CA and rate assignmentproblems To ensure network connectivity a designatedchannel is assigned to a link between every node pair Thenthe links are stored in the queue and extracted one at atime in decreasing order of priority Channels are assignedto links to minimize the maximum utilization This involvesfirst assigning the highest available transmission rate tolinks However decreasing the transmission rates on somelinks may help in improving the network performanceThis is because reducing the interference on the networkconsequently reduces the size of the linkrsquos collision domainThis is since increasing the transmission rate on a link leadsto a higher SINR-threshold at the receiver as compared withgreater difficulty in decoding of the received signals Thisalso results in more links interference with the transmissionon this link and an increase with collision domain size Theflow rates obtained from precomputed flow are then scaleddown to satisfy the sufficient condition of schedule abilitybased on the computed CA A layer-25 forwarding schemeis proposed in [70] to forward the Internet traffic to and fromgateway nodes The forwarding scheme does not maintainrouting table Instead it assigns static flow rate to its links thatforms the basis of forwarding criteria and updating of the linkcost information locally Additionally eachmesh router mustknow the hop vector to all its neighbors where a hop vectorcontains the minimum hop count to other mesh routers Thesourcemesh router inserts themaximumallowable hop countto each packet The forwarding decision is determined basedon the utilization of the links and the remaining allowable hopcount to reach its destination

In [32] the C-JCAR problem is decomposed into sub-problems wherein the number of subproblems is equal tothe number of mesh routers The subproblem at each nodeis formulated as an ILP with local constraints and similar toobjective function of [57]The nodes are grouped into subsetsdefined as network crew that is visited based on some rankingfunction For a given network crew and ranking function thenodes at each subset are visited one by one to sequentiallysolve the subproblem while the channels and paths assignedin the visited nodes are considered as constants in the nextsubproblems Branch-and-cut method is used to solve eachsubproblem Then connectivity is tested and ensured usinga postprocessing phase which also checks for unused radiointerfaces A newmetric based on contention on links routeslength and interference is used to select the best solution overall network crew and ranking

Genetic algorithm is used to model the C-JCAR problemin the network planning stage [33] For a given CA a linear


Table3State-aw

arec

hann

elassig

nmentrou

tingprop

osalssum

mary

Work

Statea

ware

Approach

Noteslimitatio

nsCh

annelreassignm

ent

Flow

rerouting

[35]

Costisb

ased

onchannel-link

reassig

nment

Not

applicablesin

ceafullyconn

ected

topo

logy

andfix

edpredefinedpaths

arec

onsid

ered

(i)Re

orderc

hann

elin

then

ewCA

toredu

cethec

ost

durin

gmigratio

nfro

mexistingclu

stertothen

ewclu

ster

(i)Out-of-o

rder

prob

lem

(ii)A

ssigning

channelsto

alllogicallin

ksaffectsthe

channeld

iversity

(iii)Perfo

rmance

degradationifthetrafficload

sequ

encesa

rehigh

lyun

correlated

[85]

Costisb

ased

onchannel-link

reassig

nment

Not

applicablesin

ceafullyconn

ected

topo

logy

isconsidered

(i)Previous

channelsarer

eassignedto

links

ifthey

area

mon

gthes

etof

channelswhich

bestredu

cethe

interfe

rencea

nddo

notviolateradioconstrains

(i)Internettraffi

cinformationiscollected

atthe

gateway

(ii)P

erform

ance

degrades

with

smalln

umbero

fradios

duetorip

plee

ffecton

channelreassignm

ent

(iii)Assigning

channelsto

alllogicallin

ksaffects

channeld

iversity

[29]

Costisb

ased

onchannel-link

reassig

nment

Not

applicablesin

ceafullyconn

ected

topo

logy

andpredefinedpathsa

reconsidered

(i)Eff

ectiv

edatafor

each

channelisc

ompu

tedand

channelw

iththeh

ighesteffectiv

ecapacity

isselected

Ifthec

hann

elisno

tpreviou

slyassig

nedto

thatlin

kreconfi

guratio

ncostissubtracted

from

thee

ffective

datavalue

(i)Out-of-o

rder

prob

lem

duetomultip

ath

(ii)A

ssigning

channelsto

alllogicallin

ksaffects

channeld

iversity

(iii)Th

echann

elassig

nedto

alinkwill

beconsidered

asfin

alassig

nmentthusearliera

ssignm

entsaren

otaw

areo

flater

assig

nments

thismay

lead

toan

improp

erfin

alchannelassignm

ent

(iv)D

ueto

radioconstraint

somelinks

will

beforced

toassig

nmorec

ongeste

dchannelsandthiswill

result

inineffi

cientchann

elassig

nment

[84]

Costisb

ased

onchannel-n

ode

reassig

nment

Not

addressed

(i)Startsfro

mthee

xisting

CAandfin

dsnewCA

bymakingsomea

djustm

ento

nthep

reviou

sone

(ii)N

umbero

fradio-chann

elreassig

nmentsis

constrainedto

apredefin

ednu

mber

(i)Perfo

rmance

degrades

with

smalln

umbero

fradios

duetorip

plee

ffecton

channelreassignm

ent

(ii)T

ransmissionrateisjointly

solved

with

CAlin

ksto

optim

izethe

networkperfo

rmance

[37]

Costisb

ased

onchannel-link

reassig

nment

Not

addressed

(i)Startin

gfro

mexistingchannelassignm

entan

ewchannelisa

ssignedto

alinkifitdo

esno

tviolatethe

radioconstraintsa

ndredu

cesthe

interfe

rence

(ii)N

umbero

flink-channelreassignm

entsis

constrainedto

apredefin

ednu

mber

(i)Perfo

rmance

degrades

with

smalln

umbero

fradio

interfa

cesa

nddu

etothatreassig

nmentisrestricted

tochannelsthatdo

notvioletthe

radioconstraints

(ii)A

ssigning

channelsto

alllogicallin

ksaffects

channeld

iversity

(iii)Flow

-levelrou

tingintro

duceso

verheadto

the

routingprocedurea

ndrequ

iresa

routingentryof

each

flowin

routingtablesTh

isisim

practic

alfor

high

numbero

fflow

s

[58]

Not

addressed

Con

sideredw

here

reroutingcostis

basedon

then

umbero

fflow

shaving

different

routingpath

(i)Heuris

ticroutingbasedon

Dijk

straa

lgorith

m(ii)Th

enum

bero

fflow

swhich

allowed

having

new

routes

isconstrainedto

apredefin

edvalue

(i)Flow

reroutingcostisno

taccurately

defin

edFor

insta

ncetwonewflo

wpath

assig

nmentsmay

result

indifferent

disrup

tionwhilebo

tharec

onsid

ered

equally

inreroutingcost


program model of multichannel routing with an objectiveof maximizing network capacity and total traffic fromto theInternet is adopted as fitness function For different hardwareresources constraints three types of optimization problemsare formulated in [33] In the first case the algorithmoptimizes the network by assigning channels to radios whileassuming the number of radios at each node is constrainedThe second case optimizes the network by assuming thatthe number of radios can be distributed into mesh routerwhere only the total number of radios in the network isconstrained while the third formulation further optimizesthe network considering the gateway placement probleminvolving selection of mesh routers to serve as gateways Inthe first case channel assigned to each radio is presented asinteger encoding with value equal to the channel assigned toit In the second and third formulations binary encoding isused for representing CA Eachmesh router is represented bybit string with a length equal to the total number of channelswith each bit representing the assignment of channel to oneof the mesh routerrsquos radios Then the algorithm starts withgeneration of an initial population The selection crossoverand mutation operation are performed at each evolutioncycle A roulette wheel procedure is used for the selectionoperation in all formulations

In [37] routing channel reassignment and TCP flowsrate allocation are studied for optimization of dynamic flowsinMR-WMN Channel reassignment is recomputed for everychange in the network traffic stateThe channel reassignmentalgorithm is used to limit the number of links whichrequires reassignment The routing model allows multipleflows between pair nodes while packets in a TCP flow followsingle path This problem is formulated as Mixed-IntegerNonlinear Program (MINLP) The objective is to maximizethe proportional fairness Heuristic algorithms are proposedto solve the C-JCAR problem The proposal starts withfinding the flow routes based on greedy heuristics The algo-rithm then selects routes iteratively based on maximum pathutilization fitness Interference-free links are assumed in therouting stage and neither radio nor interference constraint isconsidered A greedy CA routine is then carried out basedon current CA Finally flow rate allocation is performedbased on consideration of the routes and CA obtained fromprevious steps Their later work of [71] further improvestheir earlier proposal by integrating the link transmission rateallocation with the solution and by determining the activelinks to control the network topology

In [63] a concept of oblivious routing is used in solving theC-JCAR with scenarios when traffic is highly dynamic Theoblivious routing optimization aims at finding route config-uration that optimizes the worst case network performanceover a set of possible traffic demands Based on historicaltraffic demands profile with a periodic variation during thedayrsquos hours the solution starts by dividing the time intervalsinto time slots with each configuration bearing separatetime intervals The partitioning is conducted to optimize thenetwork performance by characterizing the trafficdemands ateach time slot as a convex region with number of dimensionsequal to the number of mesh routers Initially the trafficdemand is scaled such that the total traffic load from all

mesh routers at any time is constant A local search heuristicalgorithm called Hill climbing algorithm is used to partitionthe traffic profile into a predefined number of slots withthe objective of minimizing the convex hull volume overall partitions For a given time slot a traffic demand isconstructed as multidimensional convex hull with a numberof facets In case of a large number of mesh routers thenumber of facets increases exponentially Thus relaxation ofsuch convex hull to box range is constructed for such convexregion Then robust routing is computed for each timeslot The problem is mathematically formulated as nonlinearproblemwith the objective ofminimizing themaximum con-gestion over all interference sets on all channelsThe problemcomprises four sets of constraints namely interference radioflow conservation and facet constraints The formulation istransformed to LP problem by finding the dual formulationof the slave LP problem of the original problem The flow-link-channel variable is determined by solving the routingproblem using LP solver The same channel is later assignedto all links Then the algorithm iteratively selects the mostutilized link and assigns the channel obtained fromflow-link-channel variable to that flow provided that such assignmentwill not disconnect the flow Then the simulated annealingalgorithm is used to improve the solution

Some other works in the literature tackle the C-JCARproblem based on SCA configuration A heuristic approach isproposed for C-JCAR for cognitive-radio-basedWMN usingTV white space in [72]The C-JCAR is addressed with powercontrol in Liu et al [73] while [74 75] solve the C-JCARproblem for multicast traffic using genetic algorithm andgenetic algorithm with simulated annealing respectively

312 C-JCAR Based on SCA-LS In [20] the problem isformulated as an MILP and relaxed to LP with objectiveto maximize the achievable scaling factor of demands Theobtained routes from solving the LP are serving as input to theCA step The CA algorithm is performed with the objectiveof minimizing interference over all channels The obtainedchannel-link assignment ends with some infeasible links dueto the radio constraints and a postprocessing step is carriedout to redistribute the traffic load to links with feasible CAonly Thus for each logical link the total traffic load thatpassed through all channels over that link is redistributedover the common channels between the links mesh routerspairs only The objective of solving the LP formulation is tominimize the maximum interference by redistributing trafficto the links Then the scaling down of traffic demand isperformedThis ensures feasible link scheduling for realizingsufficient condition Eventually a link scheduling algorithmis also obtained

Similar to [20] in [30] the original MILP problem isrelaxed to LP with a set of four constraints The optimizationobjective maximizes the fraction of traffic load demandsthat can be routed for each node The LP is solved toobtain link-traffic-load assignments Then CA is performedto assign channels to links while the LP is solved again butwith addition of constraints to force the links to assign thesame channels obtained from the previous CA steps This isaccomplished by giving more weight (cost) to the assignment


of different channel to a link than the one that has beenpreviously assigned Finally the link scheduling is performed

In [29] CA and link scheduling are formulated as MILPproblem with the objective of maximizing throughput inCA and link scheduling The link load is first estimatedby distributing the new load of each flow on all availablepredefined paths based on hop count without consideringany interference or radio constraints A heuristic algorithmis proposed for handling the CA problem Channel reas-signment cost is considered in this stage by minimizing theamount of current traffic affected by channel reassignmentThen a greedy heuristic algorithm schedules the link to timeslots for meeting the expected load on each link Thus flowallocation is computed for determining the amount of trafficto be routed on each path with the objective of maximizingthe amount of traffic routed for each flow

313 C-JCAR Based on DCA C-JCAR are formulated as anMILP for different objective functions in [56]The problem isthen relaxed to a LP problem A prime-dual algorithm basedon shortest path routing is used to solve the LP problem byfinding the loads on links for the scaled down demands Thelink loads are then used as an input to the CA stage The CAalgorithm is developed for two scenarios the SCA-LS andDCA In SCA a greedy algorithm finds link-channel assign-ment followed by a greedy coloring algorithm for finding aconflict-free scheduling In the case of DCA packing basedheuristic algorithm is used to find both CA and schedulingsimultaneously

A C-JCAR problem with a given set of elastic flowis formulated as LP problem in [36] The objective is tomaximize the throughput using radio-to-radiomodel In thisway radio and interference constraints are replaced witha feasibility of the time fraction vector The time fractionvector represents a vector of time fraction of all links and thefraction for each linkmust be large enough to meet the trafficload requirement on that link The LP problem is solved toobtain the CA and routing Then vertex-coloring algorithmis used to obtain a feasible schedulingMore efficient coloringalgorithm leads to better throughput by lifting the resultedflows from solving the LP problem

DCAmultipath routing and link-channel scheduling arejointly formulated as MILP in [31] The objective is to maxi-mize the network fairness or throughput The work assumesthat radios can operate on subset of channels Furthermorethe possibility of combining several consecutive channelsinto one channel is explored so that network interface cardcan use the channel with larger range of frequencies Fourdifferent interference models are considered in this workThe sufficient and necessary interference constraints for loadson interfering links to be scheduled for each interferencemodel are obtained This is by deriving the interferenceconstant (which replaces 120579) for each interference modelThen mixed IP is relaxed to a linear programming by usingthe sufficient condition of the interference constraint Agreedy link scheduling algorithm is developed to find afeasible link scheduling given the flow found by the LP Thealgorithm is proved within a constant factor of the optimumsolution when it optimizes the fairness or the throughput

A parametric search is applied to improve the overallachieved flows Instead of using the sufficient condition withthe interference constant equal to one the algorithm in thefirst instance uses the maximum possible value for the inter-ference constant in the LP If no feasible solution is obtainedafter solving the LP problem and applying the schedulingalgorithm then the interference constant is reduced by oneThis is repeated until feasible solution is obtained

In [62] joint DCA and routing are considered formultipath and single-path routing under two interferencemodels The first model is the pairwise model where theinterference among the wireless links can be modeled asconflict graph This can be represented by different modelssuch as the interference-range and the protocol modelsMeanwhile the secondmodel is the physical modelThe jointproblem is formulated as LP and considered as multichan-nel multicommodity flow problem The problem considersmultipath routing with the objective of maximizing the totalthroughput of a given set of source-destination pairs withsome fairness constraints The LP is solved to obtain thetraffic load on each link The resulting flow rates are scaleddown with specific factor to obtain a feasible link schedulingThe algorithm provides a constant-approximation solutionA greedy placement approach is then used to construct linkscheduling for the scaled down links The first algorithmuses multipath routing while single-path routing presents amodified algorithm The LP problem is first solved to obtainload for each flow routed through multipath that is forcedto follow single path that is randomly chosen for each flowThen for each flow path striping is performed using theprocedure presented in [76]

A joint scheduling CA and routing is formulated as LPmulticommodity flow problem in [61] augmented with max-imal independent set (MIS) constraints First the MR-WMNresources are constructed as multidimensional conflict graph(MDCG) Each vertex inMDCG is a radio-link-channel tuple(RLC-tuple) The problem is transformed to the problem offinding all MIS in conflict graph This is because in generalsome MIS sets have more probability to be selected in theoptimal solutionThese are criticalMIS sets however findingall MIS in a graph is NP-complete problem [77] Thereforea heuristic algorithm based on scheduling index ordering isdeveloped to identify the critical MIS sets The schedulingindex metric is associated with each link that is based onthe network topology and network flow information Thealgorithm first finds all possible shortest paths between thesource-destination pairs of each flow Then the SI for eachtuple is determined based on the number of shortest pathspassing through it and its interfering neighbors The linkis then ordered based on its SI An iterative algorithm isperformed starting with tuples with higher SI first It isterminated if specific condition is reached Thus tuples forlinks associated with lower SI value will have lower possibilityto be selected The resulting maximum independent sets areused in the link capacity constraint for the multicommodityflow linear problem The solution to the linear problem isdetermined by the fraction of time where each MIS will beactive Therefore only single MIS can be active at a timeTable 2 shows a summary of the C-JCAR proposals


32 ChannelRouting Reconfiguration Proposals DCA algo-rithms assume that channel switching is supported withnegligible delay overhead [36] A per-packet switching withneglected overhead is assumed in [78] However with currenthardware channel switching delay is still considerable [79]In [80] a hybrid CA is developed and the channel switchingcost is obtained by dividing the switching delay of an interfaceover the number of packets transmitted before the interfacesswitch their channel According to [81] the delay involvedin channel switching is approximately 80 microsecondsAccording to [82] the switching delay in 80211a is around 328microseconds which is equal to the time of transmission of1024 bytes in a rate of 25Mbps using the same technologyFurthermore research [83] shows that this latency can riseup to scale of seconds due to the effect of upper layerprotocols Reference [3] presented an experimental study onthe interruption associated with channel switching which isin order of 10 seconds with the use of OLSR protocol Thisis due to fast link failure update in channel switching Thusthe authors have recommended amodification to the existingrouting protocol that can reduce the overhead associatedwith channel switching This can be concluded as even ina centralized routing paradigm that involves no link statusupdate the channel switching overhead due to synchroniza-tion issues will still be induced This is due to the inabilityof radios on a link to achieve perfectly the same switchingtime This will cause the link to be down for specific periodof time In addition to channel switching traffic reroutingcan also disrupt the real-time traffic thereby causing packetdrop The small overhead and delays caused by channelreassignment or traffic rerouting may be unacceptable forthe most of real-time multimedia applications Thereforestate-aware proposals are proposed to fill this gap There areseveral existing works which deal with channel reassignmentor traffic rerouting problems The aim of these works is toreduce the reconfiguration overhead associated with channelreassignment procedures These works are categorized asstate-aware proposals

321 Channel Reassignment Proposals Reference [79] ad-dresses the problem of recomputing CA after every vari-ation in the traffic pattern The work proposes heuristicreassignment algorithm that aims tominimize the maximumutilization in the collision domain for a given limited numberof channels The input to the algorithm is the networktopology and current CA and link-flow rates serve as inputto the heuristic algorithm The collision domain of a linkis defined as a set of links which interfere with this linkThe total utilization of a collision domain is defined as thesummation of the flow rate over all links in that collisiondomain The algorithm begins with a previous CA to findthe new one The link with the highest total utilization isselected iteratively while the channel that reduces the totalutilization over the effected collision domains is then selectedThe radio constraints may lead to inappropriate assignmentson some links due to the ripple effect and this can increasethe utilization on the affected links The number of channelreassignments is bounded by a predefined parameter Thework is further extended in [84] to integrate the transmission

rate control This is by starting with the highest possibletransmission rate to links and then decreases iteratively tillreaching the optimal configuration

Work in [85] tackles channel reassignment problem as-suming the network traffic as elastic TCP flows This algo-rithm does not account for traffic loads of each flow insteadit accounts for the number of active TCP flows This is sinceTCP flows tend to utilize the entire available bandwidth andis not restricted to a specific demand Any fluctuations in thetraffic may necessitate a consequent channel reassignment toaccount for the changes where a simple greedy algorithmhandles this problem In finding channel for a link theprevious channel on a link is reassigned if it does not violatethe radio constraints andproduces theminimum interferencelevel on that link The algorithm traverses each edge onceexcept for cases when merge operation is required and itassigns channel to it A common minimum interferencechannel is assigned to a link however if no common channelis available then merge operation is invoked to create andassign the common channel required This requires violatingthe constraints at the radio interfaces Such cases evolve insituations where the number of channels is higher than thenumber of radios on the node

Similar to [79] the channel reassignment problem is alsoconsidered in [37] also where the number of links that isallowed to reassign their channels is constrained The CAalgorithm assigns a channel which reduces the interferencein the iteration To do so the algorithm selects a channelfrom the set of channels which reduces the interference leveland do not violate the radio constraints If the candidatechannel set is empty then previously assigned channel isreassigned again If the old channel is among the candidatechannels then it is reassigned again otherwise one of thecandidate channels is assigned and the reassignment counteris incremented by one The algorithm terminates when thetotal number of reassignments is reached Meanwhile con-nectivity is maintained by assigning a channel for each logicallink The routing on TCP flow level adds extra overheadwhich usually requires complex implementation In theirtechnical report [28] the routing is inserted in each packetThus packets in each flow carry its route information tofromthe gateway while intermediate mesh routers forward eachpacket based on the included routing information Howeverinserting the full route on each packet excessively increasesthe overhead

A state-aware channel reassignment is proposed in [35]Upon receiving new traffic matrix the algorithm performs astate-aware migration to minimize the reconfiguration costsThis problem is a well-known assignment problem and it ismodeled as maximum weighted bipartite matching problemwhile thematching is determined for less reassignment trafficdisruption

Another state-aware algorithm is introduced in [29] Inthis work the CA stage calculates the expected load on eachinterference region by using the expected load on each linkwhich is then used to sort the links For each visited linkthe expected throughput is calculated for the channel andthe channel with higher expected throughput is assignedThechannel switching overhead is introduced in the expected


throughput function if the evaluated channel is different thanthe current channel on a link At any iteration unassignedlinks incident on saturated nodes are forced to assign the lastused channels that affects the channel spatial diversity

The channel switching problem in cognitive networkhas been studied in [86] As in cognitive radio networkssecondary users can operate in different licensed bands andprimary usersrsquo activities may frequently force secondaryusers to switch their radios to other unoccupied frequencybands which are unnecessarily continuous The channelswitching problem is formulated as MILP with the objectiveto reduce the frequency distance between the current and thesuccessive channels A heuristic approach is proposed to solveit However the nature of the problem is different than thechannel switching problem in the traditional MR-WMN

322 Traffic Rerouting The traffic demand dynamics cantrigger the reconfiguration of the network In this processsome flowsmight need to be reroutedThis disrupts the trafficflows and results in higher packet drop where flying packetscannot find their routes to destinations Therefore reroutingcost should be considered in the algorithm designed forenvironments with high dynamic traffic variation Manyworks avoided this problem by assuming the existence ofmultipaths between senders and receivers so that the algo-rithm maintains same paths after reconfiguration [29 35]Other works such as in [28] take the advantage of existinglinks between adjacentmesh routers to use the source routingscheme where full path to the destination is inserted in eachpacket to overcome rerouting problem

A heuristic algorithm is proposed in [58] to addressesthe flow rerouting problem The algorithm comprises threesteps a routing discovery for new flows rerouting of existingflows and assigning channels to active links The numberof rerouted flows is restricted to a predefined thresholdThe algorithm starts with discovery step to find the routesto the newly coming flows Since channel information isnot considered the traffic load on each link is distributedevenly on all channels The link cost metric for each link iscalculated and inactive links in the previous configurationare assigned with higher cost The cost for inactive linkis obtained by multiplying the normalized utilization by apredefined penalty constant Dijkstra algorithm is employedto find routes for the new flows based on the link cost In thesecond step only a predefined number of previous flows areallowed to be rerouted The new routes are selected if theyproduce less contention to the network Lastly the channelsare assigned to active links such that maximum traffic load isminimized on each channel

323 Reconfiguration Trigger In reconfiguration algorithmsthe algorithm must decide when to trigger the reconfigu-ration process Typically reconfiguration algorithms triggerthe reconfiguration process if some set of conditions aresatisfied as in [35 84 87] Reference [84] proposes to usethe maximum total utilization over all collision domains asan indicator for channel reassignment trigger In [87] thetriggering is based on traffic load monitored by the gatewayon its adjacent links The new CA is recomputed for when

the fraction of the total load on a given channel is less thana certain threshold while the reconfiguration trigger mayalso be periodic for specific interval or based on policy foraggregating similar traffic load matrices within a cluster [35]

In [35] a reconfiguration policy based on existing CA isproposed for addressing the problem of optimum utilizationand reduction of cost of reconfiguration in MR-WMNTraffic profilers at mesh nodes collect and observe the trafficinformation and provide the controller with traffic matricesperiodically Each traffic matrix presents the existing trafficdemand between node pairs The arriving traffic matricesare grouped into clusters where traffic matrices belongingto one cluster apply same CA configuration Each clusteris represented with a weighted traffic matrix of all trafficmatrices within the cluster For each newmatrix the distancebetween the new matrix and all existing clusters is foundnew traffic will then be added to the clusters if their distancesare less than a predefined value from the arriving matrixBased on the distance the algorithm determines to add thereceived traffic matrix to one of the existing clusters or toform a new cluster Channel reassignment is triggered if thereceived traffic matrix does not belong to the current clusterTable 3 shows state-aware schemes with their approaches andlimitations

To disseminate the new configuration to the other part ofthe network a dedicated radio on each mesh router is usedin [58] to exchange control information on a default channelIn [28] a simple distribution protocol is proposed This is bytaking the advantage of having a fully connected topologywhere each link is assigned to a channel A spanning treerooted at the gateway is used to disseminate the configurationwith specific control messages where higher priority is givento those messages that is by using 80211e in 80211-basedWMNs

As overall summary of the reviewed C-JCAR approachesand issues there are few points to be highlighted that canguide the network designer to develop efficient C-JCARsolutions

(i) The hardware capability of the mesh router and thewireless interface determine the type of CA schemeThe SCA-LS and DCA can achieve more efficient andflexible utilization of the radio spectrum resourcesthan SCA However DCA require a link layer syn-chronous coordination that is difficult to be achievedin multihop environment with commodity IEEE80211abg hardware Furthermore the switchingoverhead is not neglected factor and can have greatimpact over real-time traffic in multihop routesOn the other hand SCA-LS also requires link layersynchronous coordination but can be operated withlower overhead if it compared with DCA Thereforewe believe that SCA-LS is most suitable one to beimplemented

(ii) Due to scalability reasons it is not recommended tomaintain routing functionalities at layer-2 Multipathrouting is recommended to apply since it is ableto achieve load balancing and can lead to betterexploitation of the topology structure of MR-WMN


(iii) Topology is one important issue that can affect theperformance of the network Both CA and routingdetermine the physical topology of the network 1-to-1mapping as has been considered by several works tomaintain the network connectivity affects the channelspatial diversity and results in underutilizing thenetwork resources Determining the active links isone effective step for any C-JCAR algorithm

(iv) The unrealism of the current interference models canseverely affect the quality of solutions of C-JCARespecially for those applied in multihop networksTherefore it is recommended to develop a new inter-ference model based on empirical measurementThiscan be easily achievable due to the static deploy-ment of WMNs where site survey and interferenceinformation can be easily measured and recordedbeforehand

(v) Approaches in the literature still lack an accuratemodeling to the link capacity (effective capacity orvirtual capacity) of wireless link under different levelof traffic load contention interference and locationin respect to the gateway node This is essential inorder to develop efficient algorithms to optimize thetraffic distribution in multichannel multihop archi-tecture

(vi) The majority of the C-JCAR approaches in the litera-ture are load-aware and most works assume having astable traffic distribution over the network Unfortu-nately this is not the case in real scenarios where thetraffic is more dynamic since the traffic is aggregatedfrom different mesh clients with different mobilitydegrees Therefore in order to reduce the effects ofthe traffic dynamics the C-JCAR most apply adaptivestate-aware solutions and might also be integratedwith a traffic prediction model and user associationmodel to further reduce the effect of the trafficvariation on the performance

4 Practical Implementation andTest-Beds for MR-WMNs

In order to validate the optimization models for MR-WMNssimulations approaches may not lead to realistic results Thisis due to the difficulty of having an accurate characterizationof the interference in multihop environment Moreover inorder to support cross-layer solutions current hardware andprotocol stack architecture need to be revised and henceseveral practical implementations and real-life test-beds arepresented and implemented in the literature [4 88ndash95]Moreover these efforts provide the researchers with practicalframeworks to implement and evaluate their solution inreal architectures Most of these test-beds support multipleinterfaces per mesh router and are developed using off-the-shelf 80211 components or software-defined radios whenheterogeneous technologies are implemented [96 97]

Net-X project [98] is a popular framework developedto provide generic software architecture for MR-WMN Par-ticularly it is designed to support hybrid CA configuration

with two types of interfaces (I) interfaces with SCA toreceive data from neighbors and (II) switchable interfaceswith DCA to transmit data to its neighbors at configurablechannel switching intervals A channel abstraction module isimplemented between data link and network layer to supportmultiradio configuration and to coordinate the functionalityof different channels Packets are inserted to a specific channelqueue of the switchable interface to be forwarded to next-hop nodes over different channels data rates or transmittingpowers The routing table entry contains destination IPaddress next interface channel and the output interfaceA broadcast table is also maintained to select the properoutput interfaces and channels to forward the broadcastpackets

MUP (Multiradio Unification Protocol) [99] is anothersoftware architecture that supports multiradio implementa-tion using 80211-based radios Neighbor table is maintainedat link layer that stores the MAC addresses and channelinformation for each neighbor interface In [90 91] a 20-node test-bed called KAUMesh is developed in KarlstadUniversity in Sweden KAUMesh is implemented based onNet-X and OLSR [38] with a modified Network InterfaceController (NIC) Mesh routers are built by single boardplatforms and equipped with several radios based on 80211with several gateway points OLSR module is extended tosupport additional functionalities and its messages modifiedto enable exchanging the network status between meshrouters and to disseminate the channel configuration and thetraffic demands in the network

WARF [100] is a Linux based framework that designedto develop routing protocols and to perform resource mon-itoring and autonomous configuration over IPv6 networksMultiradio nodes are also supported in this architecturewhere a set of messages is defined for channel radio resourcesmonitoring and configuration This is achieved by exchang-ing control WARF messages piggybacked in the extensionheaders of IPv6 packets WARF modules can also supportcross-layer operations and different data forwarding androuting paradigm such as multipath routing

Integrating Software-Defined Networking (SDN) [101]with WMN is researched in several works [93ndash95] SDNsis an emerging architecture with a centralized controllingparadigm that enables a global control and configuration ofthe entire network through a set of configuration commandsthat are sent from SDN controllers down to the data planeThis feature enables applying global optimization strate-gies to WMNs A WMN architecture based on open-flow[102] (standard protocol in SDN) is presented in [95] Thearchitecture is used to support flow-based routing and toreduce the overhead associated with mesh clientrsquos mobilityThe architecture is implemented on KAUMesh test-bed andresults show better performance than traditional schemes

5 Conclusion and Future Research Directions

With the recent development of radio technologies nowa-days wireless technologies are expected to replace theexisting wired infrastructure in the near future In thiscontext WMNs are recognized to play a key role in the


next-generation networks Multiradio architecture based onIEEE 80211 has been introduced to improve the networkcapacity to meet these new application requirements Inthis paper an effort is made to provide an insight into theexisting approaches proposed for C-JCAR where these twoproblems are disclosed as vital issues to improve the networkperformance in multiradio architectures

The contributions of this paper are summarized as fol-lows First the key design issues for C-JCAR approaches areidentifiedThese issues cover the topology aspects algorithmdesign aspects routing and link layer specifications and traf-fic types Second the problem formulation andmathematicalmodeling are presented for the joint problem under differ-ent CA types and link specifications proposed approachesaccording to their link specifications and CA types A setof optimization objectives and fairness constraints are alsopresented Finally based on the above key designed issuesjoint configuration proposals and reconfiguration proposalsare reviewed and followed by summary tables that highlightthe design issues and present limitations and assumptionsof each work The above information can help networkoperatorplanners to decide which approaches to be usedbased on the deployment scenario network architecturehardware capabilities and traffic dynamics and characteris-tics

For the main research direction in MR-WMNs weargue that more effort needs to be done in the context ofjoint centralized approaches because joint approaches inthe literature have shown a considerable improvement tothe network performance This promotes toward integratingmultiple cross-layer solutions and optimization issues in aunified centralized framework to achieve the highest networkperformance In this architecture the optimal configurationsare obtained for all network components by taking theadvantage of having the global network view Several issuescan be addressed in this framework such as configuringthe parameters for directional antennas transmission powerand rates mac protocols channel assignments and spec-trum management routing decisions network coding andadaptive modulation usersrsquo association and mobility man-agement The framework also performs other provisioningmanagement and monitoring functions

Besides the above research direction there are still someremaining issues that still need further investigation by theC-JCAR approaches These issues include the design of anefficient channel and route reconfiguration solution with areconfiguration policy that adapts to the traffic dynamicsand the channels status with lower reconfiguration overheadEvolutionary techniques can be considered for this purposeMoreover other issues also need to be addressed in mul-tiradio architecture such as fault-management and trafficprofiling Finally developing an association mechanism formultiradio wireless mesh networks is also recommended forfurther research Thus mesh clientsrsquo association mechanismhas to be determined based on the interference and loaddistribution on backhaul links in addition to the signalstrength The association problem can also be jointly solvedwith the C-JCAR problem to further optimize the networkperformance

Notations

119873 Set of mesh routers119877 Set of radio interfacesCH Set of orthogonal channels119871 Set of links where each link 119897 = (119898 119899 119888) isin 119871

connects two mesh routers119898 and 119899 andoperates on a channel 119888

Set of links where each link = (119909 119910 119888) isin connects two radios 119909 and 119910 and operateson a channel 119888

119879 Number of time slots120579 Interference constant (based on the

interference model)119862119897 Channel capacity on link 119897

119862119874 Maximum channel capacity

119868(119897) Set of links from 119871 interfering with link 119897119892(119897) The amount of load carried on link 119897119910119905(119897) Link activation variable equaling 1 if link is

active in time slot 119905 and 0 otherwiseIR(119897) Set of links from interfering with link or

incident on one of radio pairs119905(119897) Transmitter mesh router of link 119897ℎ(119897) Receiver mesh router of link 119897119864(119899) Set of links incident in mesh router (119899)VC(119897) Virtual capacity of link 119897Ch(119897) The channel assigned to link 119897119870(119906) Number of radio interfaces in mesh router

(119906)ℎ119894

119898 Channel-node variable equal to 1 if there is

any link 119897 incident on mesh router119898 withchannel 119894 and 119892(119897) gt 0

PER(1198971| 1198972) PER at receiver node of link 119897

1when 119897

2is

active

Competing Interests

The authors declare that there is no conflict of interestsregarding the publication of this paper

References

[1] V A Siris E Z Tragos andN E Petroulakis ldquoExperiences witha metropolitan multiradio wireless mesh network design per-formance and applicationrdquo IEEE Communications Magazinevol 50 no 7 pp 128ndash136 2012

[2] M Portmann and A A Pirzada ldquoWireless mesh networks forpublic safety and crisismanagement applicationsrdquo IEEE InternetComputing vol 12 no 1 pp 18ndash25 2008

[3] S Avallone and G Di Stasi ldquoAn experimental study of thechannel switching cost in multi-radio wireless mesh networksrdquoIEEE Communications Magazine vol 51 no 9 pp 124ndash1342013

[4] I F Akyildiz XWang andWWang ldquoWirelessmesh networksa surveyrdquo Computer Networks vol 47 no 4 pp 445ndash487 2005

[5] Y Xu and W Wang ldquoWireless mesh network in smart gridmodeling and analysis for time critical communicationsrdquo IEEETransactions on Wireless Communications vol 12 no 7 pp3360ndash3371 2013


[6] P H Huynh and C E Chow ldquoDesign and analysis of hybridwireless mesh networks for smart gridsrdquo in Advances in Intel-ligent Systems and Applications-Volume 1 pp 713ndash721 SpringerNew York NY USA 2013

[7] H Son T Kang H Kim J-B Park and J H Roh ldquoA fair andsecure bandwidth allocation for AMI mesh network in smartgridrdquoThe Computer Journal vol 55 no 10 pp 1232ndash1243 2012

[8] Y Ding Y Huang G Zeng and L Xiao ldquoUsing partiallyoverlapping channels to improve throughput in wireless meshnetworksrdquo IEEE Transactions on Mobile Computing vol 11 no11 pp 1720ndash1733 2012

[9] P B FDuarte ZM Fadlullah AVVasilakos andNKato ldquoOnthe partially overlapped channel assignment on wireless meshnetwork backbone a game theoretic approachrdquo IEEE Journalon Selected Areas in Communications vol 30 no 1 pp 119ndash1272012

[10] P Cappanera L Lenzini A Lori G Stea and G VaglinildquoOptimal joint routing and link scheduling for real-time trafficin TDMAwireless mesh networksrdquo Computer Networks vol 57no 11 pp 2301ndash2312 2013

[11] M Yu X Ma W Su and L Tung ldquoA new joint strategy ofradio channel allocation and power control for wireless meshnetworksrdquo Computer Communications vol 35 no 2 pp 196ndash206 2012

[12] H Cheng and S Yang ldquoJointQoSmulticast routing and channelassignment inmultiradio multichannel wireless mesh networksusing intelligent computationalmethodsrdquoApplied Soft Comput-ing vol 11 no 2 pp 1953ndash1964 2011

[13] K Gokbayrak and E A Yıldırım ldquoJoint gateway selectiontransmission slot assignment routing and power control forwirelessmesh networksrdquoComputersampOperations Research vol40 no 7 pp 1671ndash1679 2013

[14] S Avallone ldquoAn energy efficient channel assignment androuting algorithm for multi-radio wireless mesh networksrdquo AdHoc Networks vol 10 no 6 pp 1043ndash1057 2012

[15] S Avallone F P DrsquoElia and G Ventre ldquoA new channel powerand rate assignment algorithm for multi-radio wireless meshnetworksrdquo Telecommunication Systems vol 51 no 1 pp 73ndash802012

[16] WWong X Chen F Long and S-H G Chan ldquoJoint topologycontrol and routing assignment for wireless mesh with direc-tional antennasrdquo in Proceedings of the IEEE Global Commu-nications Conference (GLOBECOM rsquo12) pp 5699ndash5704 IEEEAnaheim Calif USA December 2012

[17] A P Subramanian H Gupta S R Das and J Cao ldquoMinimuminterference channel assignment in multiradio wireless meshnetworksrdquo IEEE Transactions on Mobile Computing vol 7 no12 pp 1459ndash1473 2008

[18] K N Ramachandran E M Belding K C Almeroth andM M Buddhikot ldquoInterference-aware channel assignment inmulti-radio wireless mesh networksrdquo in Proceedings of the 25thIEEE International Conference on Computer Communications(INFOCOM rsquo06) pp 1ndash12 IEEE Barcelona Spain April 2006

[19] A Raniwala K Gopalan and T-C Chiueh ldquoCentralized chan-nel assignment and routing algorithms for multi-channel wire-less mesh networksrdquo ACM SIGMOBILE Mobile Computing andCommunications Review vol 8 pp 50ndash65 2004

[20] M Alicherry R Bhatia and L E Li ldquoJoint channel assignmentand routing for throughput optimization in multiradio wirelessmesh networksrdquo IEEE Journal on Selected Areas in Communica-tions vol 24 no 11 pp 1960ndash1971 2006

[21] B-J Ko V Misra J Padhye and D Rubenstein ldquoDistributedchannel assignment in multi-radio 80211 mesh networksrdquo inProceedings of the IEEEWireless Communications and Network-ing Conference (WCNC rsquo07) pp 3981ndash3986 Kowloon HongKong March 2007

[22] A Raniwala and T-C Chiueh ldquoArchitecture and algorithms foran IEEE 80211-based multi-channel wireless mesh networkrdquo inProceedings of the IEEE 24th Annual Joint Conference of the IEEEComputer and Communications Societies vol 3 pp 2223ndash2234IEEE March 2005

[23] J Tang G Xue and W Zhang ldquoInterference-aware topologycontrol and QoS routing in multi-channel wireless mesh net-worksrdquo in Proceedings of the 6th ACM International SymposiumonMobile Ad Hoc Networking and Computing (MOBIHOC rsquo05)pp 68ndash77 ACM Urbana-Champaign Ill USA May 2005

[24] H Skalli S Ghosh S K Das L Lenzini and M ContildquoChannel assignment strategies for multiradio wireless meshnetworks issues and solutionsrdquo IEEE Communications Maga-zine vol 45 no 11 pp 86ndash95 2007

[25] X Hong B Gu M Hoque and L Tang ldquoExploring multipleradios and multiple channels in wireless mesh networksrdquo IEEEWireless Communications vol 17 no 3 pp 76ndash85 2010

[26] PH Pathak andRDutta ldquoA survey of network design problemsand joint design approaches in wireless mesh networksrdquo IEEECommunications Surveys and Tutorials vol 13 no 3 pp 396ndash428 2011

[27] A Musaddiq F Hashim C A B C Ujang and B M AlildquoSurvey of channel assignment algorithms for multi-radiomulti-channel wireless mesh networksrdquo IETE Technical Reviewvol 32 no 3 pp 164ndash182 2015

[28] J J Galvez P M Ruiz and A Gomez-Skarmeta ResponsiveOnline Load-Balancing Routing and Load-Aware Channel Re-Assignment in Multi-Radio Multi-Channel Wireless Mesh Net-works University of Murcia Murcia Spain 2011

[29] A A Franklin A Balachandran and C S R Murthy ldquoOnlinereconfiguration of channel assignment in multi-channel multi-radio wireless mesh networksrdquo Computer Communications vol35 no 16 pp 2004ndash2013 2012

[30] M Nekoui A Ghiamatyoun S N Esfahani and M SoltanldquoIterative cross layer schemes for throughput maximization inmulti-channel wireless mesh networksrdquo in Proceedings of the16th International Conference on Computer Communicationsand Networks (ICCCN rsquo07) pp 1088ndash1092 IEEE HonoluluHawaii USA August 2007

[31] X-Y Li A Nusairat Y Wu et al ldquoJoint throughput optimiza-tion for wireless mesh networksrdquo IEEE Transactions on MobileComputing vol 8 no 7 pp 895ndash909 2009

[32] C Cicconetti V Gardellin L Lenzini and E MingozzildquoPaMeLA a joint channel assignment and routing algorithmfor multi-radio multi-channel wireless mesh networks withgrid topologyrdquo in Proceedings of the IEEE 6th InternationalConference onMobile Adhoc and Sensor Systems (MASS rsquo09) pp199ndash207 Macau China October 2009

[33] T-Y Lin K-C Hsieh and H-C Huang ldquoApplying geneticalgorithms for multiradio wireless mesh network planningrdquoIEEE Transactions on Vehicular Technology vol 61 no 5 pp2256ndash2270 2012

[34] B Claise B Trammell and P Aitken Specification of the IpFlow Information Export (Ipfix) Protocol for the Exchange of FlowInformation 2013

[35] A A Kanagasabapathy A A Franklin andC S RMurthy ldquoAnadaptive channel reconfiguration algorithm for multi-channel


multi-radio wireless mesh networksrdquo IEEE Transactions onWireless Communications vol 9 no 10 pp 3064ndash3071 2010

[36] X Meng K Tan and Q Zhang ldquoJoint routing and channelassignment in multi-radio wireless mesh networksrdquo in Proceed-ings of the IEEE International Conference on Communications(ICC rsquo06) pp 3596ndash3601 Instanbul Turkey July 2006

[37] J J Galvez and P M Ruiz ldquoEfficient rate allocation routingand channel assignment in wireless mesh networks supportingdynamic traffic flowsrdquo Ad Hoc Networks vol 11 no 6 pp 1765ndash1781 2013

[38] T Clausen and P Jacquet Optimized link state routing protocol(olsr) 2070-1721 2003

[39] Z Wang Y Chen and C Li ldquoPSR a lightweight proactivesource routing protocol for mobile ad hoc networksrdquo IEEETransactions on Vehicular Technology vol 63 no 2 pp 859ndash868 2014

[40] H Le Minh G Sexton and N Aslam ldquoSelf-adaptive proactiverouting scheme formobile ad-hoc networksrdquo IETNetworks vol4 pp 128ndash136 2014

[41] J A N Cordero J Yi and T Clausen ldquoAn adaptive jittermechanism for reactive route discovery in sensor networksrdquoSensors vol 14 no 8 pp 14440ndash14471 2014

[42] J Niu L Cheng Y Gu L Shu and S K Das ldquoR3E reliablereactive routing enhancement for wireless sensor networksrdquoIEEE Transactions on Industrial Informatics vol 10 no 1 pp784ndash794 2014

[43] S K Singh R Duvvuru and J P Singh ldquoTCP and UDP-based performance evaluation of proactive and reactive routingprotocols using mobility models in MANETSrdquo InternationalJournal of Information and Communication Technology vol 7no 6 pp 632ndash644 2015

[44] J Duan X Wang S Wang and S Xu ldquoHHR hierarchicalhybrid routing scheme for information-centric networkrdquoChinaCommunications vol 12 no 6 Article ID 7122472 pp 141ndash1532015

[45] C M S Figueiredo E F Nakamura and A A F LoureiroldquoA hybrid adaptive routing algorithm for event-driven wirelesssensor networksrdquo Sensors vol 9 no 9 pp 7287ndash7307 2009

[46] J A Guerrero Ibanez L A Garcia Morales J J ContrerasCastillo R BuenrostroMariscal andMCosio Leon ldquoHYRMAa hybrid routing protocol for monitoring of marine environ-mentsrdquo IEEELatinAmericaTransactions vol 13 no 5 pp 1562ndash1568 2015

[47] N Kojic I Reljin and B Reljin ldquoA neural networks-basedhybrid routing protocol for wireless mesh networksrdquo Sensorsvol 12 no 6 pp 7548ndash7575 2012

[48] X Chen Z Zhang and H Luo ldquoJoint optimization of powercontrol channel assignment and scheduling in wireless meshnetworkrdquo in Proceedings of the 3rd International Conference onCommunications and Networking in China (ChinaCom rsquo08) pp606ndash611 IEEE Hangzhou China August 2008

[49] A U Chaudhry N Ahmad and R H M Hafez ldquoImprovingthroughput and fairness by improved channel assignment usingtopology control based on power control for multi-radio multi-channel wireless mesh networksrdquo EURASIP Journal onWirelessCommunications and Networking vol 2012 article 155 pp 1ndash252012

[50] L Yang andLQuan ldquoA topology control algorithmusing powercontrol for wireless mesh networkrdquo in Proceedings of the 3rdInternational Conference on Multimedia Information Network-ing and Security (MINES rsquo11) pp 141ndash145 Shanghai ChinaNovember 2011

[51] Q Liu X Jia and Y Zhou ldquoTopology control for multi-channel multi-radio wireless mesh networks using directionalantennasrdquoWireless Networks vol 17 no 1 pp 41ndash51 2011

[52] A Ali M E Ahmed M J Piran and D Y Suh ldquoResourceoptimization scheme for multimedia-enabled wireless meshnetworksrdquo Sensors vol 14 no 8 pp 14500ndash14525 2014

[53] T Zhang K Yang and H-H Chen ldquoTopology control forservice-oriented wireless mesh networksrdquo IEEE Wireless Com-munications vol 16 no 4 pp 64ndash71 2009

[54] H Zhang and D H K Tsang ldquoTraffic oriented topology for-mation and load-balancing routing in wireless mesh networksrdquoin Proceedings of the 16th International Conference on ComputerCommunications and Networks (ICCCN rsquo07) pp 1046ndash1052IEEE Honolulu Hawaii USA August 2007

[55] F Liu and Y Bai ldquoAn overview of topology control mecha-nisms in multi-radio multi-channel wireless mesh networksrdquoEURASIP Journal onWireless Communications and Networkingvol 2012 article 324 2012

[56] M Kodialam and T Nandagopal ldquoCharacterizing the capacityregion in multi-radio multi-channel wireless mesh networksrdquoin Proceedings of the 11th Annual International Conference onMobile Computing and Networking (MobiCom rsquo05) pp 73ndash87ACM Cologne Germany September 2005

[57] A H Mohsenian-Rad and V W S Wong ldquoJoint logicaltopology design interface assignment channel allocation androuting formulti-channel wirelessmesh networksrdquo IEEE Trans-actions on Wireless Communications vol 6 no 12 pp 4432ndash4440 2007

[58] Y Zhou S-H Chung and H-Y Jeong ldquoReconfiguring multi-rate wi-fi mesh networks with flow disruption constraintsrdquo inDynamics in Logistics pp 383ndash393 Springer New York NYUSA 2013

[59] A Iyer C Rosenberg and A Karnik ldquoWhat is the rightmodel for wireless channel interferencerdquo IEEE Transactions onWireless Communications vol 8 no 5 pp 2662ndash2671 2009

[60] A U Chaudhry R H M Hafez and J W Chinneck ldquoOn theimpact of interference models on channel assignment in multi-radiomulti-channel wirelessmesh networksrdquoAdHocNetworksvol 27 pp 68ndash80 2015

[61] H Li Y Cheng C Zhou and P Wan ldquoMulti-dimensionalconflict graph based computing for optimal capacity inMR-MCwireless networksrdquo in Proceedings of the 30th IEEE InternationalConference on Distributed Computing Systems (ICDCS rsquo10) pp774ndash783 IEEE Genova Italy June 2010

[62] M Al-Ayyoub and H Gupta ldquoJoint routing channel assign-ment and scheduling for throughput maximization in generalinterference modelsrdquo IEEE Transactions on Mobile Computingvol 9 no 4 pp 553ndash565 2010

[63] J Wellons and Y Xue ldquoThe robust joint solution for channelassignment and routing for wireless mesh networks with timepartitioningrdquo Ad Hoc Networks vol 13 pp 210ndash221 2014

[64] V Gardellin S K Das L Lenzini C Cicconetti and E Min-gozzi ldquoG-PaMeLA a divide-and-conquer approach for jointchannel assignment and routing in multi-radio multi-channelwireless mesh networksrdquo Journal of Parallel and DistributedComputing vol 71 no 3 pp 381ndash396 2011

[65] O Goussevskaia Y A Oswald and R Wattenhofer ldquoComplex-ity in geometric SINRrdquo in Proceedings of the 8th ACM Interna-tional Symposium onMobile AdHoc Networking and Computing(MobiHoc rsquo07) pp 100ndash109Montreal Canada September 2007

[66] V Kumar M V Marathe S Parthasarathy and A SrinivasanldquoEnd-to-end packet-scheduling in wireless ad-hoc networksrdquo


in Proceedings of the 15th Annual ACM-SIAM Symposium onDiscrete Algorithms (SODA rsquo04) pp 1021ndash1030 Society forIndustrial and Applied Mathematics 2004

[67] CPLEXV12 1 UserrsquosManual for CPLEX International BusinessMachines Corporation 2009

[68] X Huang S-L Feng and H-C Zhuang ldquoCross-layer fairresources allocation for multi-radio multi-channel wirelessmesh networksrdquo in Proceedings of the 5th International Confer-ence onWireless Communications Networking andMobile Com-puting (WiCOM rsquo09) pp 1ndash5 IEEE Beijing China September2009

[69] N Sadeghianpour T C Chuah and S W Tan ldquoJoint channelassignment and routing in multiradio multichannel wirelessmesh networkswith directional antennasrdquo International Journalof Communication Systems vol 28 no 9 pp 1521ndash1536 2015

[70] S Avallone I F Akyildiz and G Ventre ldquoA channel and rateassignment algorithm and a layer-25 forwarding paradigm formulti-radio wireless mesh networksrdquo IEEEACM Transactionson Networking vol 17 no 1 pp 267ndash280 2009

[71] J J Galvez and P M Ruiz ldquoJoint link rate allocation routingand channel assignment in multi-rate multi-channel wirelessnetworksrdquo Ad Hoc Networks vol 29 pp 78ndash98 2015

[72] R Kikuchi K Hosaki andM Hasegawa ldquoImplementation andevaluation of a combined optimization scheme for routing andchannel assignment in wireless mesh networksrdquo in Proceedingsof the 12th Annual IEEE Consumer Communications and Net-working Conference (CCNC rsquo15) pp 619ndash624 IEEE Las VegasNev USA January 2015

[73] K-M Liu T Ma Y-A Liu and K-H Kou ldquoFairness-orientedrouting algorithm joint with power control and channel assign-ment for multi-radio multi-channel wireless mesh networksrdquoThe Journal of China Universities of Posts and Telecommunica-tions vol 21 no 5 pp 55ndash60 2014

[74] M Sharifi Jebeli and M Dehghan ldquoJoint multicast routing andchannel assignment in Multiradio Multichannel Wireless MeshNetworks using a multi objective algorithmrdquo in Proceedings ofthe 6th Conference on Information and Knowledge Technology(IKT rsquo14) pp 163ndash170 IEEE Shahrud Iran May 2014

[75] E Vaezpour and M Dehghan ldquoA multi-objective optimizationapproach for joint channel assignment and multicast routingin multi-radio multi-channel wireless mesh networksrdquoWirelessPersonal Communications vol 77 no 2 pp 1055ndash1076 2014

[76] P Raghavan and C D Tompson ldquoRandomized roundinga technique for provably good algorithms and algorithmicproofsrdquo Combinatorica vol 7 no 4 pp 365ndash374 1987

[77] M R Garey and D S Johnson Computers and IntractabilityA Guide to the Theory of NP-Completeness Freeman SanFrancisco La USA 1979

[78] X Shao C Hua and A Huang ldquoRobust resource allocationfor multi-hop wireless mesh networks with end-to-end trafficspecificationsrdquo Ad Hoc Networks vol 13 pp 123ndash133 2014

[79] S Avallone F P DrsquoElia and G Ventre ldquoA traffic-aware channelre-assignment algorithm for wireless mesh networksrdquo in Pro-ceedings of the European Wireless Conference (EW rsquo10) pp 683ndash688 IEEE Lucca Italy April 2010

[80] P Kyasanur and N H Vaidya ldquoRouting and interface assign-ment in multi-channel multi-interface wireless networksrdquo inProceedings of the IEEEWireless Communications and Network-ing Conference (WCNC rsquo05) pp 2051ndash2056 IEEE NewOrleansLa USA March 2005

[81] F Herzel G Fischer and H Gustat ldquoAn integrated CMOS RFsynthesizer for 80211a wireless LANrdquo IEEE Journal of Solid-State Circuits vol 38 no 10 pp 1767ndash1770 2003

[82] M Yun Y Zhou A Arora and H-A Choi ldquoChannel-assign-ment and scheduling in wireless mesh networks consideringswitching overheadrdquo in Proceedings of the IEEE InternationalConference on Communications (ICC rsquo09) pp 1ndash6 IEEE Dres-den Germany June 2009

[83] P Li N Scalabrino Y Fang E Gregori and I Chlamtac ldquoHowto effectively use multiple channels in wireless mesh networksrdquoIEEE Transactions on Parallel and Distributed Systems vol 20no 11 pp 1641ndash1652 2009

[84] S Avallone G Di Stasi and A Kassler ldquoA traffic-aware channeland rate reassignment algorithm for wireless mesh networksrdquoIEEE Transactions onMobile Computing vol 12 no 7 pp 1335ndash1348 2013

[85] J J Galvez P M Ruiz and A F G Skarmeta ldquoTCP flow-awarechannel re-assignment in multi-radio multi-channel wirelessmesh networksrdquo in Proceedings of the 8th IEEE InternationalConference onMobile Ad-hoc and Sensor Systems (MASS rsquo11) pp262ndash271 Valencia Spain October 2011

[86] S Arkoulis E Anifantis V Karyotis S Papavassiliou and NMitrou ldquoOn the optimal fair and channel-aware cognitive radionetwork reconfigurationrdquoComputer Networks vol 57 no 8 pp1739ndash1757 2013

[87] W Fu B Xie XWang and D P Agrawal ldquoFlow-based channelassignment in channel constrained wireless mesh networksrdquo inProceedings of the 17th International Conference on ComputerCommunications and Networks (ICCCN rsquo08) pp 1ndash6 IEEE StThomas Virgin Islands USA August 2008

[88] P Dely M Castro S Soukhakian A Moldsvor and A KasslerldquoPractical considerations for channel assignment in wirelessmesh networksrdquo inProceedings of the IEEEGlobecomWorkshops(GC Wkshps rsquo10) pp 763ndash767 IEEE Miami Fla USA Decem-ber 2010

[89] B Blywis M Guenes F Juraschek and J H Schiller ldquoTrendsadvances and challenges in testbed-based wireless mesh net-work researchrdquoMobile Networks and Applications vol 15 no 3pp 315ndash329 2010

[90] P Dely and A Kassler ldquoKaumesh a multi-radio multi-channelmesh testbedrdquo in Proceedings of 9th Scandinavian Workshop onWireless Ad-hocampSensorNetworks (ADHOC rsquo09) Niagara FallsCanada 2009

[91] M C Castro P Dely A J Kassler F P DrsquoElia and S AvalloneldquoOlsr and net-X as a framework for channel assignment exper-imentsrdquo in Proceedings of the 4th ACM International Workshopon Experimental Evaluation and Characterization (WINTECHrsquo09) pp 79ndash80 ACM Beijing China 2009

[92] G Di Stasi R Bifulco S Avallone et al ldquoInterconnection ofgeographically distributed wireless mesh testbeds resourcesharing on a large scalerdquoAdHocNetworks vol 9 no 8 pp 1389ndash1403 2011

[93] D Zhu X Yang P Zhao and W Yu ldquoTowards effective intra-flow network coding in software defined wireless mesh net-worksrdquo in Proceedings of the 24th International Conference onComputer Communication and Networks (ICCCN rsquo15) pp 1ndash8IEEE Las Vegas Nev USA 2015

[94] P Dely ldquoTowards an architecture for openflow and wirelessmesh networksrdquo in Proceedings of the Ofelia Summer SchoolNovember 2011

[95] P Dely A Kassler and N Bayer ldquoOpenFlow for wireless meshnetworksrdquo inProceedings of the 20th International Conference on


Computer Communications and Networks (ICCCN rsquo11) pp 1ndash6Maui Hawaii USA August 2011

[96] V Angelakis A Capone A Fragkiadakis et al ldquoExperiencefrom testbeds and management platforms towards mesh net-working with heterogeneous wireless accessrdquo in Proceedings ofthe IEEE 14th International Symposium and Workshops on aWorld ofWirelessMobile andMultimediaNetworks (WoWMoMrsquo13) pp 1ndash6 Madrid Spain June 2013

[97] Redcomm project April 2016 httpwwwredcomm-projecteudescriptionhtml

[98] P Kyasanur C Chereddi and N H Vaidya ldquoNet-x systemextensions for supporting multiple channels multiple inter-faces and other interface capabilitiesrdquo Tech Rep Universityof Illinois at Urbana-Champaign Wireless Networking GroupUrbana Ill USA 2006

[99] A Adya P Bahl J Padhye A Wolman and L Zhou ldquoA multi-radio unification protocol for IEEE 80211 wireless networksrdquoin Proceedings of the 1st International Conference on BroadbandNetworks (BroadNets rsquo04) pp 344ndash354 IEEE San Jose CalifUSA October 2004

[100] S Kuklinski P Radziszewski and J Wytrębowicz ldquoWARFcomponent based platform for wireless mesh networksrdquo SmartCR vol 1 no 2 pp 125ndash138 2011

[101] S Ortiz Jr ldquoSoftware-defined networking on the verge of abreakthroughrdquo IEEE Computer vol 46 no 7 pp 10ndash12 2013

[102] N McKeown T Anderson H Balakrishnan et al ldquoOpenFlowenabling innovation in campus networksrdquo ACM SIGCOMMComputer Communication Review vol 38 no 2 pp 69ndash742008

International Journal of

AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

RoboticsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014


Active and Passive Electronic Components

Control Scienceand Engineering

Journal of



RotatingMachinery


Hindawi Publishing Corporation httpwwwhindawicom

Journal ofEngineeringVolume 2014

Submit your manuscripts athttpwwwhindawicom

VLSI Design



Shock and Vibration


Civil EngineeringAdvances in

Acoustics and VibrationAdvances in



Electrical and Computer Engineering

Journal of

Advances inOptoElectronics


Volume 2014

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

SensorsJournal of


Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014


Chemical EngineeringInternational Journal of Antennas and

Propagation




Navigation and Observation



DistributedSensor Networks



Multiradio mesh router

Mesh clientBackhaul wireless linkAccess wireless link

Wired link

Internet

CH3

CH3

CH3

CH3

CH3

CH6

CH6

CH6

CH6

CH6CH8

CH8

CH8CH8

CH8CH8

CH8

CH8

CH8

Gateway

Gateway

Figure 1 The multiradio wireless mesh networks architecture

Recently MR-WMN has attracted numerous numbersof research efforts to utilize the advantages that this net-work is offering Several proposed approaches on channelassignment (CA) algorithms multichannel MAC protocolsmultichannel routing metrics links scheduling (LS) mul-tichannel multicast protocols power and topology controland network planning exist in the literature However thedesigns that are considering combinations of these issueshave shown more efficient performance such as joint routingand link scheduling [10] joint CA and power control [11]joint QoS multicast routing and CA [12] joint gatewayselection transmission slot assignment routing and powercontrol [13] joint CA power control and routing [14] jointCA power control and rate assignment [15] joint routing andtopology control with directional antennas [16] and partiallyoverlapped CA [8 9]

CA algorithms aim to assign channels to the radiointerfaces and links with the objective of minimizing theoverall contention and interference over the wireless linksCA can either be solved as a separated problem as in [17 18] orjointly solved with routing as in [19 20] Furthermore it canbe developed as centralized solutions [19 20] and distributedsolutions [21 22] The availability of the entire network viewmakes centralized solution more effective than distributedsolutions which only relayed on local information TypicallyCA algorithms must have knowledge on network load inorder to assign channels to links However some otheralgorithms do not require any traffic information as in [23]where the interference is minimized by conserving a 119896-connected topology

Joint design approaches in WMNs are surveyed in manyworks in the literature such as [24ndash27] However none ofthe works in the literature have comprehensively investigatedthe Centralized Joint Channel Assignment and Routing (C-JCAR) approaches This motivates us to undertake an in-depth review of different C-JCAR proposals in the litera-ture Figure 2 shows the research direction of this paperThe rest of the paper is organized as follows Section 2discusses key design issues models and approaches for C-JCAR algorithms Section 3 presents reviewed works undertwo subsections of C-JCAR approaches and reconfigurationCArouting algorithms Section 4 investigates some prac-tical implantations and test-beds for MR-WMNs FinallySection 5 describes the future research directions and con-cludes the paper

2 Design Issues Modeling andApproaches for C-JCAR

This section identifies the key design issues from link and net-work layers for C-JCAR approaches to address and clarifiesinterferencemodel and networkmodel to be used in problemformulation and algorithms It also presents themathematicalformulation of the problem including optimization objec-tives constraints and fairness Besides this section also givesthe general classifications to C-JCAR algorithms

21 Key Design Issues211 Channel Assignment Schemes Based on the underlinghardware CA can be implemented in one of three schemes



1

2

3


Related works


Design issues

Modeling

Approaches





Reconfiguration

C-JCAR









SCA

SCA-LS

DCA Time(slotted)

Time(slotted)

Time

Ch1

Ch1 Ch1Ch8 Ch8

Ch1

Ch1

Ch1

Ch5






119879)



119868) where 119877

119868=




|V minus 119909| le 119877119879(1)




(a) 1-to-1 mapping






Transmission range

Wireless link

z

n

m

y

u

x

(a)

Interference range

Interferingm

n

z

y

u

x

transmission

(b)





|V minus 119909| le 119877119879

(2)



119875V

119873 + sum119895isinradiosV 119875119895

ge 120573V119909 (3)




n

m

z

y

u

x

| minus x||nminusx|

|uminusx|



(a)




n

m

y

z

u

x

u y z n m

(b)



Interference range

z

n

m

y

u

x














Directional link


Radio interface




64]





0119891ge 120582 lowast 120596



120583 ge0max0min

(5)



maximize sum

119891isinFLlog (0

119891) (6)















Link capacity (119897) =119905119897

119879sdot 119862119897 (7)






VC (119897) =119892 (119897)

sum1198971015840isin119868(119897)cup119897

119892 (1198971015840)sdot 119862119874 119897 isin 119871 (8)


VC (119897) + sum

1198971015840isin119868(119897)

VC (1198971015840) = 119862119874

forall1198971015840isin 119871 (9)


119892 (119897)

VC (119897)le 1 forall119897 isin 119871 (10)



sum

1198971015840isin119868(119897)cup119897

119892 (1198971015840)

1198621198971015840

le 120579 forall119897 isin 119871 (11)




sum

1198971015840isin119868(119897)cup119897

119892 (1198971015840)

1198621198971015840

le 1 forall119897 isin 119871 (12)


Time slots


Wireless link

Link

s e2

e5

e1

e4e3

e2(03)

e5(04)e1(02)

e4(03)

e3(04)





Minimize sum

1198971

sum

1198972isin119868(1198971)

119892 (1198971) lowast 119892 (119897


1| 1198972) (13)




SCA


general


sum

1198971015840isin119868(119897)cup119897

119892 (1198971015840)

1198621198971015840

le 120579 forall119897 isin 119871


sum

1198971015840isin119868(119897)

119892 (1198971015840)

1198621198971015840


IPsum

119894isin119862

ℎ119894

119898le 119870 (119906) forall119906 isin 119873

LPsum

119897

ℎ(119897)=119898

or 119905(119897)=119898

119892 (119897)

119862119897

le 119870 (119906) forall119906 isin 119873


[37]


1198971015840isin119868(119897)cup119897


119892 (119897)

VC (119897)le 1 forall119897 isin 119871

DCASCA-LS


general




1198971015840isin119868(119897)cup119897



sum

1198971015840isin119868(119897)cup119897

119892 (1198971015840)

1198621198971015840

le 120579 forall119897 isin 119871


sum

1198971015840isin119868(119897)

119892 (1198971015840)

1198621198971015840


IPsum

119897isin119871

119905(119897)=119906

or ℎ(119897)=119906

119910119905(119897) le 119870(119906) forall119906 isin 119873

LPsum

119897isin119871

119905(119897)=119906

or ℎ(119897)=119906

119892 (119897)

119862119897

le 119870 (119906) forall119906 isin 119873


[56]


sum

1198971015840isin

119864(119899)cup119864(119898)



sum

1198971015840isin119868(119897)

119892 (1198971015840)

1198621198971015840




119905(1198971015840

) le 1 forall1198971015840

isin 119868 (119897)


sum

1015840

isin119868(119897)

119892 (1198971015840

)

1198621015840


IPsum

isin119871

119905(119897)=119906

or ℎ(119897)=119906

119910119905(119897) le 119870 (119906) forall119906 isin 119873

LP

sum

isin

119905(119897)=119903

or ℎ(119897)=119903

119892 (119897)

119862




sum

1015840

isinIR(119897)cup119897

119892 (1198971015840

)

1198621015840





Connectivity

Link mapping

Topology info

Channelassignment

Interferencemodel

Technology

Traffic type

Traffic profile

Traffic collection

Routing scheme

Routing table

Execution (Re-)



Algorithms
















Topology

Approaches Traffic

Routing layer

Link layer





1when



119895 Let 119860 isin 119860

119895 Δ(119860) be the set


sum

119890isinΔ(119860)

119909 (119897) le 120593 forall119895 119860 isin 119860119895 (14)






1198821198972

1198971

= 120573 lowast119875V (119903)

119875V (119906)

forall1198971 1198972isin 119871 amp 119897

1= (119903 119904 119888) 1198972 = (119906 V 119888)

(15)



119892 (1198971) + sum

1198972isin1198711198971ampCh(119897

2)=Ch(119897

1)

119873(1198821198972

1198971

) sdot 119892 (1198972) le 120599 (119897

1)


(16)



3 Related Works



Table2C-

JCARprop

osalssum

mary

Work

Link

layera

ndchannelassignm

ent

Traffi

camprouting

Topo

logy

Objectiv

eampfairn

ess

Noteslimitatio

nsCA

type

Interfe

rence

Techno

logy

Traffi

ctype

Traffi

cload

Routingscheme

Mapping

(1-to-119909)119909

Con

nectivity

Objectiv

eFairn

ess

[19]

SCA

B-IRM

80211

P2P

Yes

SPM

P1

Yes

Maxim

izec

ross-sectio

ngood

-put

No

Optim

ality

isno


kcapacityestim

ationin

aheavy

traffi

cloadLess

channeld

iversitydu

etolin

kmapping

[56]

DCA

SC

A-LS

B-IRM

mdashP2

PYes

MP

01

Yes

Maxim

ize120582

(scalin

gfactor)

Min-120582

Out-of-o

rder

prob

lem

andconfi

guratio

ncomplexity

onroutingSw

itching

overhead

issues

[20]

SCA-

LSB-IRM

mdashInternet

Yes

MP

Any

ForM

Rswtr

affic

Maxim

ize120582

(scalin

gfactor)

Min-120582

Radioconstraintsa

reno

tcon

sidered

intheL

POut-of-o

rder

prob

lem

duetomultip

athrouting

[36]

DCA

B-IRM

mdashP2

Psrc-desM

RsMP

Any

No

Maxim

izea

ggregated

throug

hput

No

Out-of-o

rder

prob

lem

andconfi

guratio

ncomplexity

onroutingSw

itching

overhead

issues

Optim

ality

issue

since

onlysufficientcon

ditio

nsarec

onsid

ered

[57]

SCA

B-IRM

80211

P2PInternet

Yes

SPAny

ForM

Rswtr

affic

Maxim

izem

insparelink

capacity

Min-120582

onlin

kutilizatio

n

Usin

gsufficientcon

ditio

nin

MILPmay

notlead

toglob

alop

timalsolutio

nSing

lechannelinitia

ltopo

logy

may

lead

toinfeasiblesolutio

nin

heavy

traffi

cloadscenarios

[30]

SCA-

LSTH

MInternet

Yes

MPpredefined

01

ForM

Rswtr

affic

Maxim

ize120582

(scalin

gfactor)

Min-120582

Con

sideringon

lynecessarycond

ition

may

lead

toinfeasiblesolutio

nOut-of-o

rder

prob

lem

dueto

multip

athrouting

[70]

SCA

PHM

(wo

accumulation)

80211

Internet

No

MP(M

-GW)

L25forw

arding

ge1

Yes

Minim

izem

axlin

kutilizatio

nNo

Radioor

interfe

rencec

onstr

aintsa

reno

taccoun

tedin

initialtraffi

cloadestim

ation

Assum

esam

etrafficd

emands

inallm

eshrouters

Out-of-o

rder

prob

lem

duetomultip

athrouting

Link

rateassig

nmentisc

onsid

ered

[31]

DCA

PRMIRM

Internet

Yes

MP(M

-GW)

Any

ForM

Rswtr

affic

Maxim

izea

ggregated

throug

hput

Min-120582

Out-of-o

rder

prob

lem

andconfi

guratio

ncomplexity

onroutingEa

chradiocanbe

operated

onas

ubseto

fchann

elsC

ombinatio

nof

multip

lechannelsto

singlec

hann

elispo

ssible

[32]

SCA

THM

Internet

Yes

SPAny

Yes

Metric

onho

pcoun

tampinterfe

rence

No

Solvingsubp

roblem

with

localcon

straintsdo

esno

tguarantee

glob

alop


received

power

atan

odew

henanothern

odeis

transm

ittingisassumed

available

[62]

DCA

IRMP

RM

PHM

P2P

No

SPM

PAny

No

Maxim

izea

ggregated

throug

hput

Max-to

-Min


atmessrou

tersIn

single-path

routingop

timality

isno

tguaranteed

since

mathematicalform

ulationisform

ultip

ath

[61]

DCA

B-IRM

P2P

Yes

MP

Any

No

Maxim

ize120582

(scalin

gfactor)

Min-120582

Theq

ualityof

thes

olutionishigh

lydepend

ento

nthep

erform

ance

ofcriticalM

ISfin

ding

algorithm

[33]

SCA

B-IRM

80211

Internet

Lower

ampup

per

boun

dMP(M

GW)

Any

ForM

Rswtr

affic

Maxim

izea

ggregated

throug

hput

Boun

dedfairn

ess

Develo

pedforp

lann

ingsta

gewith

different

resourcesc

onstr

aintsNot

practic

alforlarge

networks

since

fitnessfunctio

ninvolves

solving

entiren

etworkLP

[29]

SCA-

LSTH

MP2

PYes

MP(predefin

ed)

1Yes

Maxim

izea

ggrachieved

demands

No

Link

loadsu

sedto

theC

Aaree

stimated

with

out

consideringradioor

interfe

rencec

onstr

aintsLess

channeld

iversitydu

etolin

kmapping


Table2Con

tinued

Work

Link

layera

ndchannelassignm

ent

Traffi

camprouting

Topo

logy

Objectiv

eampfairn

ess

Noteslimitatio

nsCA

type

Interfe

rence

Techno

logy

Traffi

ctype

Traffi

cload

Routingscheme

Mapping

(1-to-119909)119909

Con

nectivity

Objectiv

eFairn

ess

[37]

SCA

EMInternet

Num

bero

factiv

eflow

sMP(SPper

follo

w)

1Yes

Minim

izec

ontention

(CA)

No

Link

loadse

stimated

with

outcon

sideringradioor

interfe

rencec

onstr

aintsNot

clearlysta

tedwhere

andho

wto

implem

entfl

owratecontrol

[63]

SCA

B-IRM

80211

Internet

Yessequ

ence

oftraffi

cMP

Any

ForM

Rswtr

affic

Minim

izethe

maxim

umcongestio

n(ofw

orst

case)

No

Num

bero

ffacetsincreasee

xpon

entia

llyin

large-scalen


then

umbero

fconstraintsinLP

prob

lemhow

everrelaxingthe

convex

hullto

boxrang

esaffectsthee

fficiency

NoteSCA

static

channelassignm

entDCA

dyn

amicchannelassignm

entLS

link

schedu

lingIM

interfe

rencem

odel

IRMin

terfe

rence-rangem

odel

PRMprotocolm

odel

PHMphysic

almod

elB-bidire

ctional

THMtwo-ho

pmod

elEM

empiric

almeasurementP2

Ppeer-to

-peerSP

single-pathM

Pmultip

athMGWm

ultig

atew

ayM

Rmeshrouterand

LLlinklayer



31 C-JCAR Proposals








Table3State-aw

arec

hann

elassig

nmentrou

tingprop

osalssum

mary

Work

Statea

ware

Approach

Noteslimitatio

nsCh

annelreassignm

ent

Flow

rerouting

[35]

Costisb

ased

onchannel-link

reassig

nment

Not

applicablesin

ceafullyconn

ected

topo

logy

andfix

edpredefinedpaths

arec

onsid

ered

(i)Re

orderc

hann

elin

then

ewCA

toredu

cethec

ost

durin

gmigratio

nfro

mexistingclu

stertothen

ewclu

ster

(i)Out-of-o

rder

prob

lem

(ii)A

ssigning

channelsto

alllogicallin

ksaffectsthe

channeld

iversity

(iii)Perfo

rmance


sequ

encesa

rehigh

lyun

correlated

[85]

Costisb

ased

onchannel-link

reassig

nment

Not

applicablesin

ceafullyconn

ected

topo

logy

isconsidered

(i)Previous

channelsarer

eassignedto

links

ifthey

area

mon

gthes

etof

channelswhich

bestredu

cethe

interfe

rencea

nddo


(i)Internettraffi


atthe

gateway

(ii)P

erform

ance

degrades

with

smalln

umbero

fradios

duetorip

plee

ffecton

channelreassignm

ent

(iii)Assigning

channelsto

alllogicallin

ksaffects

channeld

iversity

[29]

Costisb

ased

onchannel-link

reassig

nment

Not

applicablesin

ceafullyconn

ected

topo

logy

andpredefinedpathsa

reconsidered

(i)Eff

ectiv

edatafor

each

channelisc

ompu

tedand

channelw

iththeh

ighesteffectiv

ecapacity

isselected

Ifthec

hann

elisno

tpreviou

slyassig

nedto

thatlin

kreconfi

guratio

ncostissubtracted

from

thee

ffective

datavalue

(i)Out-of-o

rder

prob

lem

duetomultip

ath

(ii)A

ssigning

channelsto

alllogicallin

ksaffects

channeld

iversity

(iii)Th

echann

elassig

nedto

alinkwill

beconsidered

asfin

alassig

nmentthusearliera

ssignm

entsaren

otaw

areo

flater

assig

nments

thismay

lead

toan

improp

erfin

alchannelassignm

ent

(iv)D

ueto

radioconstraint

somelinks

will

beforced

toassig

nmorec

ongeste


result

inineffi

cientchann

elassig

nment

[84]

Costisb

ased

onchannel-n

ode

reassig

nment

Not

addressed

(i)Startsfro

mthee

xisting

CAandfin

dsnewCA

bymakingsomea

djustm

ento

nthep

reviou

sone

(ii)N

umbero

fradio-chann

elreassig

nmentsis

constrainedto

apredefin

ednu

mber

(i)Perfo

rmance

degrades

with

smalln

umbero

fradios

duetorip

plee

ffecton

channelreassignm

ent

(ii)T


solved

with

CAlin

ksto

optim

izethe

networkperfo

rmance

[37]

Costisb

ased

onchannel-link

reassig

nment

Not

addressed

(i)Startin

gfro


entan

ewchannelisa

ssignedto

alinkifitdo

esno

tviolatethe

radioconstraintsa

ndredu

cesthe

interfe

rence

(ii)N

umbero


entsis

constrainedto

apredefin

ednu

mber

(i)Perfo

rmance

degrades

with

smalln

umbero

fradio

interfa

cesa

nddu

etothatreassig

nmentisrestricted

tochannelsthatdo

notvioletthe

radioconstraints

(ii)A

ssigning

channelsto

alllogicallin

ksaffects

channeld

iversity

(iii)Flow

-levelrou

tingintro

duceso

verheadto

the

routingprocedurea

ndrequ

iresa

routingentryof

each

flowin

routingtablesTh

isisim

practic

alfor

high

numbero

fflow

s

[58]

Not

addressed

Con

sideredw

here

reroutingcostis

basedon

then

umbero

fflow

shaving

different

routingpath

(i)Heuris

ticroutingbasedon

Dijk

straa

lgorith

m(ii)Th

enum

bero

fflow

swhich

allowed

having

new

routes

isconstrainedto

apredefin

edvalue

(i)Flow

reroutingcostisno

taccurately

defin

edFor

insta

ncetwonewflo

wpath

assig

nmentsmay

result

indifferent

disrup

tionwhilebo

tharec

onsid

ered

equally

inreroutingcost

























































Notations










(119906)ℎ119894




1when 119897

2is

active

Competing Interests


References
















































































































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation











1

2

3


Related works


Design issues

Modeling

Approaches





Reconfiguration

C-JCAR









SCA

SCA-LS

DCA Time(slotted)

Time(slotted)

Time

Ch1

Ch1 Ch1Ch8 Ch8

Ch1

Ch1

Ch1

Ch5






119879)



119868) where 119877

119868=




|V minus 119909| le 119877119879(1)




(a) 1-to-1 mapping






Transmission range

Wireless link

z

n

m

y

u

x

(a)

Interference range

Interferingm

n

z

y

u

x

transmission

(b)





|V minus 119909| le 119877119879

(2)



119875V

119873 + sum119895isinradiosV 119875119895

ge 120573V119909 (3)




n

m

z

y

u

x

| minus x||nminusx|

|uminusx|



(a)




n

m

y

z

u

x

u y z n m

(b)



Interference range

z

n

m

y

u

x














Directional link


Radio interface




64]





0119891ge 120582 lowast 120596



120583 ge0max0min

(5)



maximize sum

119891isinFLlog (0

119891) (6)















Link capacity (119897) =119905119897

119879sdot 119862119897 (7)






VC (119897) =119892 (119897)

sum1198971015840isin119868(119897)cup119897

119892 (1198971015840)sdot 119862119874 119897 isin 119871 (8)


VC (119897) + sum

1198971015840isin119868(119897)

VC (1198971015840) = 119862119874

forall1198971015840isin 119871 (9)


119892 (119897)

VC (119897)le 1 forall119897 isin 119871 (10)



sum

1198971015840isin119868(119897)cup119897

119892 (1198971015840)

1198621198971015840

le 120579 forall119897 isin 119871 (11)




sum

1198971015840isin119868(119897)cup119897

119892 (1198971015840)

1198621198971015840

le 1 forall119897 isin 119871 (12)


Time slots


Wireless link

Link

s e2

e5

e1

e4e3

e2(03)

e5(04)e1(02)

e4(03)

e3(04)





Minimize sum

1198971

sum

1198972isin119868(1198971)

119892 (1198971) lowast 119892 (119897


1| 1198972) (13)




SCA


general


sum

1198971015840isin119868(119897)cup119897

119892 (1198971015840)

1198621198971015840

le 120579 forall119897 isin 119871


sum

1198971015840isin119868(119897)

119892 (1198971015840)

1198621198971015840


IPsum

119894isin119862

ℎ119894

119898le 119870 (119906) forall119906 isin 119873

LPsum

119897

ℎ(119897)=119898

or 119905(119897)=119898

119892 (119897)

119862119897

le 119870 (119906) forall119906 isin 119873


[37]


1198971015840isin119868(119897)cup119897


119892 (119897)

VC (119897)le 1 forall119897 isin 119871

DCASCA-LS


general




1198971015840isin119868(119897)cup119897



sum

1198971015840isin119868(119897)cup119897

119892 (1198971015840)

1198621198971015840

le 120579 forall119897 isin 119871


sum

1198971015840isin119868(119897)

119892 (1198971015840)

1198621198971015840


IPsum

119897isin119871

119905(119897)=119906

or ℎ(119897)=119906

119910119905(119897) le 119870(119906) forall119906 isin 119873

LPsum

119897isin119871

119905(119897)=119906

or ℎ(119897)=119906

119892 (119897)

119862119897

le 119870 (119906) forall119906 isin 119873


[56]


sum

1198971015840isin

119864(119899)cup119864(119898)



sum

1198971015840isin119868(119897)

119892 (1198971015840)

1198621198971015840




119905(1198971015840

) le 1 forall1198971015840

isin 119868 (119897)


sum

1015840

isin119868(119897)

119892 (1198971015840

)

1198621015840


IPsum

isin119871

119905(119897)=119906

or ℎ(119897)=119906

119910119905(119897) le 119870 (119906) forall119906 isin 119873

LP

sum

isin

119905(119897)=119903

or ℎ(119897)=119903

119892 (119897)

119862




sum

1015840

isinIR(119897)cup119897

119892 (1198971015840

)

1198621015840





Connectivity

Link mapping

Topology info

Channelassignment

Interferencemodel

Technology

Traffic type

Traffic profile

Traffic collection

Routing scheme

Routing table

Execution (Re-)



Algorithms
















Topology

Approaches Traffic

Routing layer

Link layer





1when



119895 Let 119860 isin 119860

119895 Δ(119860) be the set


sum

119890isinΔ(119860)

119909 (119897) le 120593 forall119895 119860 isin 119860119895 (14)






1198821198972

1198971

= 120573 lowast119875V (119903)

119875V (119906)

forall1198971 1198972isin 119871 amp 119897

1= (119903 119904 119888) 1198972 = (119906 V 119888)

(15)



119892 (1198971) + sum

1198972isin1198711198971ampCh(119897

2)=Ch(119897

1)

119873(1198821198972

1198971

) sdot 119892 (1198972) le 120599 (119897

1)


(16)



3 Related Works



Table2C-

JCARprop

osalssum

mary

Work

Link

layera

ndchannelassignm

ent

Traffi

camprouting

Topo

logy

Objectiv

eampfairn

ess

Noteslimitatio

nsCA

type

Interfe

rence

Techno

logy

Traffi

ctype

Traffi

cload

Routingscheme

Mapping

(1-to-119909)119909

Con

nectivity

Objectiv

eFairn

ess

[19]

SCA

B-IRM

80211

P2P

Yes

SPM

P1

Yes

Maxim

izec

ross-sectio

ngood

-put

No

Optim

ality

isno


kcapacityestim

ationin

aheavy

traffi

cloadLess

channeld

iversitydu

etolin

kmapping

[56]

DCA

SC

A-LS

B-IRM

mdashP2

PYes

MP

01

Yes

Maxim

ize120582

(scalin

gfactor)

Min-120582

Out-of-o

rder

prob

lem

andconfi

guratio

ncomplexity

onroutingSw

itching

overhead

issues

[20]

SCA-

LSB-IRM

mdashInternet

Yes

MP

Any

ForM

Rswtr

affic

Maxim

ize120582

(scalin

gfactor)

Min-120582

Radioconstraintsa

reno

tcon

sidered

intheL

POut-of-o

rder

prob

lem

duetomultip

athrouting

[36]

DCA

B-IRM

mdashP2

Psrc-desM

RsMP

Any

No

Maxim

izea

ggregated

throug

hput

No

Out-of-o

rder

prob

lem

andconfi

guratio

ncomplexity

onroutingSw

itching

overhead

issues

Optim

ality

issue

since

onlysufficientcon

ditio

nsarec

onsid

ered

[57]

SCA

B-IRM

80211

P2PInternet

Yes

SPAny

ForM

Rswtr

affic

Maxim

izem

insparelink

capacity

Min-120582

onlin

kutilizatio

n

Usin

gsufficientcon

ditio

nin

MILPmay

notlead

toglob

alop

timalsolutio

nSing

lechannelinitia

ltopo

logy

may

lead

toinfeasiblesolutio

nin

heavy

traffi

cloadscenarios

[30]

SCA-

LSTH

MInternet

Yes

MPpredefined

01

ForM

Rswtr

affic

Maxim

ize120582

(scalin

gfactor)

Min-120582

Con

sideringon

lynecessarycond

ition

may

lead

toinfeasiblesolutio

nOut-of-o

rder

prob

lem

dueto

multip

athrouting

[70]

SCA

PHM

(wo

accumulation)

80211

Internet

No

MP(M

-GW)

L25forw

arding

ge1

Yes

Minim

izem

axlin

kutilizatio

nNo

Radioor

interfe

rencec

onstr

aintsa

reno

taccoun

tedin

initialtraffi

cloadestim

ation

Assum

esam

etrafficd

emands

inallm

eshrouters

Out-of-o

rder

prob

lem

duetomultip

athrouting

Link

rateassig

nmentisc

onsid

ered

[31]

DCA

PRMIRM

Internet

Yes

MP(M

-GW)

Any

ForM

Rswtr

affic

Maxim

izea

ggregated

throug

hput

Min-120582

Out-of-o

rder

prob

lem

andconfi

guratio

ncomplexity

onroutingEa

chradiocanbe

operated

onas

ubseto

fchann

elsC

ombinatio

nof

multip

lechannelsto

singlec

hann

elispo

ssible

[32]

SCA

THM

Internet

Yes

SPAny

Yes

Metric

onho

pcoun

tampinterfe

rence

No

Solvingsubp

roblem

with

localcon

straintsdo

esno

tguarantee

glob

alop


received

power

atan

odew

henanothern

odeis

transm

ittingisassumed

available

[62]

DCA

IRMP

RM

PHM

P2P

No

SPM

PAny

No

Maxim

izea

ggregated

throug

hput

Max-to

-Min


atmessrou

tersIn

single-path

routingop

timality

isno

tguaranteed

since

mathematicalform

ulationisform

ultip

ath

[61]

DCA

B-IRM

P2P

Yes

MP

Any

No

Maxim

ize120582

(scalin

gfactor)

Min-120582

Theq

ualityof

thes

olutionishigh

lydepend

ento

nthep

erform

ance

ofcriticalM

ISfin

ding

algorithm

[33]

SCA

B-IRM

80211

Internet

Lower

ampup

per

boun

dMP(M

GW)

Any

ForM

Rswtr

affic

Maxim

izea

ggregated

throug

hput

Boun

dedfairn

ess

Develo

pedforp

lann

ingsta

gewith

different

resourcesc

onstr

aintsNot

practic

alforlarge

networks

since

fitnessfunctio

ninvolves

solving

entiren

etworkLP

[29]

SCA-

LSTH

MP2

PYes

MP(predefin

ed)

1Yes

Maxim

izea

ggrachieved

demands

No

Link

loadsu

sedto

theC

Aaree

stimated

with

out

consideringradioor

interfe

rencec

onstr

aintsLess

channeld

iversitydu

etolin

kmapping


Table2Con

tinued

Work

Link

layera

ndchannelassignm

ent

Traffi

camprouting

Topo

logy

Objectiv

eampfairn

ess

Noteslimitatio

nsCA

type

Interfe

rence

Techno

logy

Traffi

ctype

Traffi

cload

Routingscheme

Mapping

(1-to-119909)119909

Con

nectivity

Objectiv

eFairn

ess

[37]

SCA

EMInternet

Num

bero

factiv

eflow

sMP(SPper

follo

w)

1Yes

Minim

izec

ontention

(CA)

No

Link

loadse

stimated

with

outcon

sideringradioor

interfe

rencec

onstr

aintsNot

clearlysta

tedwhere

andho

wto

implem

entfl

owratecontrol

[63]

SCA

B-IRM

80211

Internet

Yessequ

ence

oftraffi

cMP

Any

ForM

Rswtr

affic

Minim

izethe

maxim

umcongestio

n(ofw

orst

case)

No

Num

bero

ffacetsincreasee

xpon

entia

llyin

large-scalen


then

umbero

fconstraintsinLP

prob

lemhow

everrelaxingthe

convex

hullto

boxrang

esaffectsthee

fficiency

NoteSCA

static

channelassignm

entDCA

dyn

amicchannelassignm

entLS

link

schedu

lingIM

interfe

rencem

odel

IRMin

terfe

rence-rangem

odel

PRMprotocolm

odel

PHMphysic

almod

elB-bidire

ctional

THMtwo-ho

pmod

elEM

empiric

almeasurementP2

Ppeer-to

-peerSP

single-pathM

Pmultip

athMGWm

ultig

atew

ayM

Rmeshrouterand

LLlinklayer



31 C-JCAR Proposals








Table3State-aw

arec

hann

elassig

nmentrou

tingprop

osalssum

mary

Work

Statea

ware

Approach

Noteslimitatio

nsCh

annelreassignm

ent

Flow

rerouting

[35]

Costisb

ased

onchannel-link

reassig

nment

Not

applicablesin

ceafullyconn

ected

topo

logy

andfix

edpredefinedpaths

arec

onsid

ered

(i)Re

orderc

hann

elin

then

ewCA

toredu

cethec

ost

durin

gmigratio

nfro

mexistingclu

stertothen

ewclu

ster

(i)Out-of-o

rder

prob

lem

(ii)A

ssigning

channelsto

alllogicallin

ksaffectsthe

channeld

iversity

(iii)Perfo

rmance


sequ

encesa

rehigh

lyun

correlated

[85]

Costisb

ased

onchannel-link

reassig

nment

Not

applicablesin

ceafullyconn

ected

topo

logy

isconsidered

(i)Previous

channelsarer

eassignedto

links

ifthey

area

mon

gthes

etof

channelswhich

bestredu

cethe

interfe

rencea

nddo


(i)Internettraffi


atthe

gateway

(ii)P

erform

ance

degrades

with

smalln

umbero

fradios

duetorip

plee

ffecton

channelreassignm

ent

(iii)Assigning

channelsto

alllogicallin

ksaffects

channeld

iversity

[29]

Costisb

ased

onchannel-link

reassig

nment

Not

applicablesin

ceafullyconn

ected

topo

logy

andpredefinedpathsa

reconsidered

(i)Eff

ectiv

edatafor

each

channelisc

ompu

tedand

channelw

iththeh

ighesteffectiv

ecapacity

isselected

Ifthec

hann

elisno

tpreviou

slyassig

nedto

thatlin

kreconfi

guratio

ncostissubtracted

from

thee

ffective

datavalue

(i)Out-of-o

rder

prob

lem

duetomultip

ath

(ii)A

ssigning

channelsto

alllogicallin

ksaffects

channeld

iversity

(iii)Th

echann

elassig

nedto

alinkwill

beconsidered

asfin

alassig

nmentthusearliera

ssignm

entsaren

otaw

areo

flater

assig

nments

thismay

lead

toan

improp

erfin

alchannelassignm

ent

(iv)D

ueto

radioconstraint

somelinks

will

beforced

toassig

nmorec

ongeste


result

inineffi

cientchann

elassig

nment

[84]

Costisb

ased

onchannel-n

ode

reassig

nment

Not

addressed

(i)Startsfro

mthee

xisting

CAandfin

dsnewCA

bymakingsomea

djustm

ento

nthep

reviou

sone

(ii)N

umbero

fradio-chann

elreassig

nmentsis

constrainedto

apredefin

ednu

mber

(i)Perfo

rmance

degrades

with

smalln

umbero

fradios

duetorip

plee

ffecton

channelreassignm

ent

(ii)T


solved

with

CAlin

ksto

optim

izethe

networkperfo

rmance

[37]

Costisb

ased

onchannel-link

reassig

nment

Not

addressed

(i)Startin

gfro


entan

ewchannelisa

ssignedto

alinkifitdo

esno

tviolatethe

radioconstraintsa

ndredu

cesthe

interfe

rence

(ii)N

umbero


entsis

constrainedto

apredefin

ednu

mber

(i)Perfo

rmance

degrades

with

smalln

umbero

fradio

interfa

cesa

nddu

etothatreassig

nmentisrestricted

tochannelsthatdo

notvioletthe

radioconstraints

(ii)A

ssigning

channelsto

alllogicallin

ksaffects

channeld

iversity

(iii)Flow

-levelrou

tingintro

duceso

verheadto

the

routingprocedurea

ndrequ

iresa

routingentryof

each

flowin

routingtablesTh

isisim

practic

alfor

high

numbero

fflow

s

[58]

Not

addressed

Con

sideredw

here

reroutingcostis

basedon

then

umbero

fflow

shaving

different

routingpath

(i)Heuris

ticroutingbasedon

Dijk

straa

lgorith

m(ii)Th

enum

bero

fflow

swhich

allowed

having

new

routes

isconstrainedto

apredefin

edvalue

(i)Flow

reroutingcostisno

taccurately

defin

edFor

insta

ncetwonewflo

wpath

assig

nmentsmay

result

indifferent

disrup

tionwhilebo

tharec

onsid

ered

equally

inreroutingcost

























































Notations










(119906)ℎ119894




1when 119897

2is

active

Competing Interests


References
















































































































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation










SCA

SCA-LS

DCA Time(slotted)

Time(slotted)

Time

Ch1

Ch1 Ch1Ch8 Ch8

Ch1

Ch1

Ch1

Ch5






119879)



119868) where 119877

119868=




|V minus 119909| le 119877119879(1)




(a) 1-to-1 mapping






Transmission range

Wireless link

z

n

m

y

u

x

(a)

Interference range

Interferingm

n

z

y

u

x

transmission

(b)





|V minus 119909| le 119877119879

(2)



119875V

119873 + sum119895isinradiosV 119875119895

ge 120573V119909 (3)




n

m

z

y

u

x

| minus x||nminusx|

|uminusx|



(a)




n

m

y

z

u

x

u y z n m

(b)



Interference range

z

n

m

y

u

x














Directional link


Radio interface




64]





0119891ge 120582 lowast 120596



120583 ge0max0min

(5)



maximize sum

119891isinFLlog (0

119891) (6)















Link capacity (119897) =119905119897

119879sdot 119862119897 (7)






VC (119897) =119892 (119897)

sum1198971015840isin119868(119897)cup119897

119892 (1198971015840)sdot 119862119874 119897 isin 119871 (8)


VC (119897) + sum

1198971015840isin119868(119897)

VC (1198971015840) = 119862119874

forall1198971015840isin 119871 (9)


119892 (119897)

VC (119897)le 1 forall119897 isin 119871 (10)



sum

1198971015840isin119868(119897)cup119897

119892 (1198971015840)

1198621198971015840

le 120579 forall119897 isin 119871 (11)




sum

1198971015840isin119868(119897)cup119897

119892 (1198971015840)

1198621198971015840

le 1 forall119897 isin 119871 (12)


Time slots


Wireless link

Link

s e2

e5

e1

e4e3

e2(03)

e5(04)e1(02)

e4(03)

e3(04)





Minimize sum

1198971

sum

1198972isin119868(1198971)

119892 (1198971) lowast 119892 (119897


1| 1198972) (13)




SCA


general


sum

1198971015840isin119868(119897)cup119897

119892 (1198971015840)

1198621198971015840

le 120579 forall119897 isin 119871


sum

1198971015840isin119868(119897)

119892 (1198971015840)

1198621198971015840


IPsum

119894isin119862

ℎ119894

119898le 119870 (119906) forall119906 isin 119873

LPsum

119897

ℎ(119897)=119898

or 119905(119897)=119898

119892 (119897)

119862119897

le 119870 (119906) forall119906 isin 119873


[37]


1198971015840isin119868(119897)cup119897


119892 (119897)

VC (119897)le 1 forall119897 isin 119871

DCASCA-LS


general




1198971015840isin119868(119897)cup119897



sum

1198971015840isin119868(119897)cup119897

119892 (1198971015840)

1198621198971015840

le 120579 forall119897 isin 119871


sum

1198971015840isin119868(119897)

119892 (1198971015840)

1198621198971015840


IPsum

119897isin119871

119905(119897)=119906

or ℎ(119897)=119906

119910119905(119897) le 119870(119906) forall119906 isin 119873

LPsum

119897isin119871

119905(119897)=119906

or ℎ(119897)=119906

119892 (119897)

119862119897

le 119870 (119906) forall119906 isin 119873


[56]


sum

1198971015840isin

119864(119899)cup119864(119898)



sum

1198971015840isin119868(119897)

119892 (1198971015840)

1198621198971015840




119905(1198971015840

) le 1 forall1198971015840

isin 119868 (119897)


sum

1015840

isin119868(119897)

119892 (1198971015840

)

1198621015840


IPsum

isin119871

119905(119897)=119906

or ℎ(119897)=119906

119910119905(119897) le 119870 (119906) forall119906 isin 119873

LP

sum

isin

119905(119897)=119903

or ℎ(119897)=119903

119892 (119897)

119862




sum

1015840

isinIR(119897)cup119897

119892 (1198971015840

)

1198621015840





Connectivity

Link mapping

Topology info

Channelassignment

Interferencemodel

Technology

Traffic type

Traffic profile

Traffic collection

Routing scheme

Routing table

Execution (Re-)



Algorithms
















Topology

Approaches Traffic

Routing layer

Link layer





1when



119895 Let 119860 isin 119860

119895 Δ(119860) be the set


sum

119890isinΔ(119860)

119909 (119897) le 120593 forall119895 119860 isin 119860119895 (14)






1198821198972

1198971

= 120573 lowast119875V (119903)

119875V (119906)

forall1198971 1198972isin 119871 amp 119897

1= (119903 119904 119888) 1198972 = (119906 V 119888)

(15)



119892 (1198971) + sum

1198972isin1198711198971ampCh(119897

2)=Ch(119897

1)

119873(1198821198972

1198971

) sdot 119892 (1198972) le 120599 (119897

1)


(16)



3 Related Works



Table2C-

JCARprop

osalssum

mary

Work

Link

layera

ndchannelassignm

ent

Traffi

camprouting

Topo

logy

Objectiv

eampfairn

ess

Noteslimitatio

nsCA

type

Interfe

rence

Techno

logy

Traffi

ctype

Traffi

cload

Routingscheme

Mapping

(1-to-119909)119909

Con

nectivity

Objectiv

eFairn

ess

[19]

SCA

B-IRM

80211

P2P

Yes

SPM

P1

Yes

Maxim

izec

ross-sectio

ngood

-put

No

Optim

ality

isno


kcapacityestim

ationin

aheavy

traffi

cloadLess

channeld

iversitydu

etolin

kmapping

[56]

DCA

SC

A-LS

B-IRM

mdashP2

PYes

MP

01

Yes

Maxim

ize120582

(scalin

gfactor)

Min-120582

Out-of-o

rder

prob

lem

andconfi

guratio

ncomplexity

onroutingSw

itching

overhead

issues

[20]

SCA-

LSB-IRM

mdashInternet

Yes

MP

Any

ForM

Rswtr

affic

Maxim

ize120582

(scalin

gfactor)

Min-120582

Radioconstraintsa

reno

tcon

sidered

intheL

POut-of-o

rder

prob

lem

duetomultip

athrouting

[36]

DCA

B-IRM

mdashP2

Psrc-desM

RsMP

Any

No

Maxim

izea

ggregated

throug

hput

No

Out-of-o

rder

prob

lem

andconfi

guratio

ncomplexity

onroutingSw

itching

overhead

issues

Optim

ality

issue

since

onlysufficientcon

ditio

nsarec

onsid

ered

[57]

SCA

B-IRM

80211

P2PInternet

Yes

SPAny

ForM

Rswtr

affic

Maxim

izem

insparelink

capacity

Min-120582

onlin

kutilizatio

n

Usin

gsufficientcon

ditio

nin

MILPmay

notlead

toglob

alop

timalsolutio

nSing

lechannelinitia

ltopo

logy

may

lead

toinfeasiblesolutio

nin

heavy

traffi

cloadscenarios

[30]

SCA-

LSTH

MInternet

Yes

MPpredefined

01

ForM

Rswtr

affic

Maxim

ize120582

(scalin

gfactor)

Min-120582

Con

sideringon

lynecessarycond

ition

may

lead

toinfeasiblesolutio

nOut-of-o

rder

prob

lem

dueto

multip

athrouting

[70]

SCA

PHM

(wo

accumulation)

80211

Internet

No

MP(M

-GW)

L25forw

arding

ge1

Yes

Minim

izem

axlin

kutilizatio

nNo

Radioor

interfe

rencec

onstr

aintsa

reno

taccoun

tedin

initialtraffi

cloadestim

ation

Assum

esam

etrafficd

emands

inallm

eshrouters

Out-of-o

rder

prob

lem

duetomultip

athrouting

Link

rateassig

nmentisc

onsid

ered

[31]

DCA

PRMIRM

Internet

Yes

MP(M

-GW)

Any

ForM

Rswtr

affic

Maxim

izea

ggregated

throug

hput

Min-120582

Out-of-o

rder

prob

lem

andconfi

guratio

ncomplexity

onroutingEa

chradiocanbe

operated

onas

ubseto

fchann

elsC

ombinatio

nof

multip

lechannelsto

singlec

hann

elispo

ssible

[32]

SCA

THM

Internet

Yes

SPAny

Yes

Metric

onho

pcoun

tampinterfe

rence

No

Solvingsubp

roblem

with

localcon

straintsdo

esno

tguarantee

glob

alop


received

power

atan

odew

henanothern

odeis

transm

ittingisassumed

available

[62]

DCA

IRMP

RM

PHM

P2P

No

SPM

PAny

No

Maxim

izea

ggregated

throug

hput

Max-to

-Min


atmessrou

tersIn

single-path

routingop

timality

isno

tguaranteed

since

mathematicalform

ulationisform

ultip

ath

[61]

DCA

B-IRM

P2P

Yes

MP

Any

No

Maxim

ize120582

(scalin

gfactor)

Min-120582

Theq

ualityof

thes

olutionishigh

lydepend

ento

nthep

erform

ance

ofcriticalM

ISfin

ding

algorithm

[33]

SCA

B-IRM

80211

Internet

Lower

ampup

per

boun

dMP(M

GW)

Any

ForM

Rswtr

affic

Maxim

izea

ggregated

throug

hput

Boun

dedfairn

ess

Develo

pedforp

lann

ingsta

gewith

different

resourcesc

onstr

aintsNot

practic

alforlarge

networks

since

fitnessfunctio

ninvolves

solving

entiren

etworkLP

[29]

SCA-

LSTH

MP2

PYes

MP(predefin

ed)

1Yes

Maxim

izea

ggrachieved

demands

No

Link

loadsu

sedto

theC

Aaree

stimated

with

out

consideringradioor

interfe

rencec

onstr

aintsLess

channeld

iversitydu

etolin

kmapping


Table2Con

tinued

Work

Link

layera

ndchannelassignm

ent

Traffi

camprouting

Topo

logy

Objectiv

eampfairn

ess

Noteslimitatio

nsCA

type

Interfe

rence

Techno

logy

Traffi

ctype

Traffi

cload

Routingscheme

Mapping

(1-to-119909)119909

Con

nectivity

Objectiv

eFairn

ess

[37]

SCA

EMInternet

Num

bero

factiv

eflow

sMP(SPper

follo

w)

1Yes

Minim

izec

ontention

(CA)

No

Link

loadse

stimated

with

outcon

sideringradioor

interfe

rencec

onstr

aintsNot

clearlysta

tedwhere

andho

wto

implem

entfl

owratecontrol

[63]

SCA

B-IRM

80211

Internet

Yessequ

ence

oftraffi

cMP

Any

ForM

Rswtr

affic

Minim

izethe

maxim

umcongestio

n(ofw

orst

case)

No

Num

bero

ffacetsincreasee

xpon

entia

llyin

large-scalen


then

umbero

fconstraintsinLP

prob

lemhow

everrelaxingthe

convex

hullto

boxrang

esaffectsthee

fficiency

NoteSCA

static

channelassignm

entDCA

dyn

amicchannelassignm

entLS

link

schedu

lingIM

interfe

rencem

odel

IRMin

terfe

rence-rangem

odel

PRMprotocolm

odel

PHMphysic

almod

elB-bidire

ctional

THMtwo-ho

pmod

elEM

empiric

almeasurementP2

Ppeer-to

-peerSP

single-pathM

Pmultip

athMGWm

ultig

atew

ayM

Rmeshrouterand

LLlinklayer



31 C-JCAR Proposals








Table3State-aw

arec

hann

elassig

nmentrou

tingprop

osalssum

mary

Work

Statea

ware

Approach

Noteslimitatio

nsCh

annelreassignm

ent

Flow

rerouting

[35]

Costisb

ased

onchannel-link

reassig

nment

Not

applicablesin

ceafullyconn

ected

topo

logy

andfix

edpredefinedpaths

arec

onsid

ered

(i)Re

orderc

hann

elin

then

ewCA

toredu

cethec

ost

durin

gmigratio

nfro

mexistingclu

stertothen

ewclu

ster

(i)Out-of-o

rder

prob

lem

(ii)A

ssigning

channelsto

alllogicallin

ksaffectsthe

channeld

iversity

(iii)Perfo

rmance


sequ

encesa

rehigh

lyun

correlated

[85]

Costisb

ased

onchannel-link

reassig

nment

Not

applicablesin

ceafullyconn

ected

topo

logy

isconsidered

(i)Previous

channelsarer

eassignedto

links

ifthey

area

mon

gthes

etof

channelswhich

bestredu

cethe

interfe

rencea

nddo


(i)Internettraffi


atthe

gateway

(ii)P

erform

ance

degrades

with

smalln

umbero

fradios

duetorip

plee

ffecton

channelreassignm

ent

(iii)Assigning

channelsto

alllogicallin

ksaffects

channeld

iversity

[29]

Costisb

ased

onchannel-link

reassig

nment

Not

applicablesin

ceafullyconn

ected

topo

logy

andpredefinedpathsa

reconsidered

(i)Eff

ectiv

edatafor

each

channelisc

ompu

tedand

channelw

iththeh

ighesteffectiv

ecapacity

isselected

Ifthec

hann

elisno

tpreviou

slyassig

nedto

thatlin

kreconfi

guratio

ncostissubtracted

from

thee

ffective

datavalue

(i)Out-of-o

rder

prob

lem

duetomultip

ath

(ii)A

ssigning

channelsto

alllogicallin

ksaffects

channeld

iversity

(iii)Th

echann

elassig

nedto

alinkwill

beconsidered

asfin

alassig

nmentthusearliera

ssignm

entsaren

otaw

areo

flater

assig

nments

thismay

lead

toan

improp

erfin

alchannelassignm

ent

(iv)D

ueto

radioconstraint

somelinks

will

beforced

toassig

nmorec

ongeste


result

inineffi

cientchann

elassig

nment

[84]

Costisb

ased

onchannel-n

ode

reassig

nment

Not

addressed

(i)Startsfro

mthee

xisting

CAandfin

dsnewCA

bymakingsomea

djustm

ento

nthep

reviou

sone

(ii)N

umbero

fradio-chann

elreassig

nmentsis

constrainedto

apredefin

ednu

mber

(i)Perfo

rmance

degrades

with

smalln

umbero

fradios

duetorip

plee

ffecton

channelreassignm

ent

(ii)T


solved

with

CAlin

ksto

optim

izethe

networkperfo

rmance

[37]

Costisb

ased

onchannel-link

reassig

nment

Not

addressed

(i)Startin

gfro


entan

ewchannelisa

ssignedto

alinkifitdo

esno

tviolatethe

radioconstraintsa

ndredu

cesthe

interfe

rence

(ii)N

umbero


entsis

constrainedto

apredefin

ednu

mber

(i)Perfo

rmance

degrades

with

smalln

umbero

fradio

interfa

cesa

nddu

etothatreassig

nmentisrestricted

tochannelsthatdo

notvioletthe

radioconstraints

(ii)A

ssigning

channelsto

alllogicallin

ksaffects

channeld

iversity

(iii)Flow

-levelrou

tingintro

duceso

verheadto

the

routingprocedurea

ndrequ

iresa

routingentryof

each

flowin

routingtablesTh

isisim

practic

alfor

high

numbero

fflow

s

[58]

Not

addressed

Con

sideredw

here

reroutingcostis

basedon

then

umbero

fflow

shaving

different

routingpath

(i)Heuris

ticroutingbasedon

Dijk

straa

lgorith

m(ii)Th

enum

bero

fflow

swhich

allowed

having

new

routes

isconstrainedto

apredefin

edvalue

(i)Flow

reroutingcostisno

taccurately

defin

edFor

insta

ncetwonewflo

wpath

assig

nmentsmay

result

indifferent

disrup

tionwhilebo

tharec

onsid

ered

equally

inreroutingcost

























































Notations










(119906)ℎ119894




1when 119897

2is

active

Competing Interests


References
















































































































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation












(a) 1-to-1 mapping






Transmission range

Wireless link

z

n

m

y

u

x

(a)

Interference range

Interferingm

n

z

y

u

x

transmission

(b)





|V minus 119909| le 119877119879

(2)



119875V

119873 + sum119895isinradiosV 119875119895

ge 120573V119909 (3)




n

m

z

y

u

x

| minus x||nminusx|

|uminusx|



(a)




n

m

y

z

u

x

u y z n m

(b)



Interference range

z

n

m

y

u

x














Directional link


Radio interface




64]





0119891ge 120582 lowast 120596



120583 ge0max0min

(5)



maximize sum

119891isinFLlog (0

119891) (6)















Link capacity (119897) =119905119897

119879sdot 119862119897 (7)






VC (119897) =119892 (119897)

sum1198971015840isin119868(119897)cup119897

119892 (1198971015840)sdot 119862119874 119897 isin 119871 (8)


VC (119897) + sum

1198971015840isin119868(119897)

VC (1198971015840) = 119862119874

forall1198971015840isin 119871 (9)


119892 (119897)

VC (119897)le 1 forall119897 isin 119871 (10)



sum

1198971015840isin119868(119897)cup119897

119892 (1198971015840)

1198621198971015840

le 120579 forall119897 isin 119871 (11)




sum

1198971015840isin119868(119897)cup119897

119892 (1198971015840)

1198621198971015840

le 1 forall119897 isin 119871 (12)


Time slots


Wireless link

Link

s e2

e5

e1

e4e3

e2(03)

e5(04)e1(02)

e4(03)

e3(04)





Minimize sum

1198971

sum

1198972isin119868(1198971)

119892 (1198971) lowast 119892 (119897


1| 1198972) (13)




SCA


general


sum

1198971015840isin119868(119897)cup119897

119892 (1198971015840)

1198621198971015840

le 120579 forall119897 isin 119871


sum

1198971015840isin119868(119897)

119892 (1198971015840)

1198621198971015840


IPsum

119894isin119862

ℎ119894

119898le 119870 (119906) forall119906 isin 119873

LPsum

119897

ℎ(119897)=119898

or 119905(119897)=119898

119892 (119897)

119862119897

le 119870 (119906) forall119906 isin 119873


[37]


1198971015840isin119868(119897)cup119897


119892 (119897)

VC (119897)le 1 forall119897 isin 119871

DCASCA-LS


general




1198971015840isin119868(119897)cup119897



sum

1198971015840isin119868(119897)cup119897

119892 (1198971015840)

1198621198971015840

le 120579 forall119897 isin 119871


sum

1198971015840isin119868(119897)

119892 (1198971015840)

1198621198971015840


IPsum

119897isin119871

119905(119897)=119906

or ℎ(119897)=119906

119910119905(119897) le 119870(119906) forall119906 isin 119873

LPsum

119897isin119871

119905(119897)=119906

or ℎ(119897)=119906

119892 (119897)

119862119897

le 119870 (119906) forall119906 isin 119873


[56]


sum

1198971015840isin

119864(119899)cup119864(119898)



sum

1198971015840isin119868(119897)

119892 (1198971015840)

1198621198971015840




119905(1198971015840

) le 1 forall1198971015840

isin 119868 (119897)


sum

1015840

isin119868(119897)

119892 (1198971015840

)

1198621015840


IPsum

isin119871

119905(119897)=119906

or ℎ(119897)=119906

119910119905(119897) le 119870 (119906) forall119906 isin 119873

LP

sum

isin

119905(119897)=119903

or ℎ(119897)=119903

119892 (119897)

119862




sum

1015840

isinIR(119897)cup119897

119892 (1198971015840

)

1198621015840





Connectivity

Link mapping

Topology info

Channelassignment

Interferencemodel

Technology

Traffic type

Traffic profile

Traffic collection

Routing scheme

Routing table

Execution (Re-)



Algorithms
















Topology

Approaches Traffic

Routing layer

Link layer





1when



119895 Let 119860 isin 119860

119895 Δ(119860) be the set


sum

119890isinΔ(119860)

119909 (119897) le 120593 forall119895 119860 isin 119860119895 (14)






1198821198972

1198971

= 120573 lowast119875V (119903)

119875V (119906)

forall1198971 1198972isin 119871 amp 119897

1= (119903 119904 119888) 1198972 = (119906 V 119888)

(15)



119892 (1198971) + sum

1198972isin1198711198971ampCh(119897

2)=Ch(119897

1)

119873(1198821198972

1198971

) sdot 119892 (1198972) le 120599 (119897

1)


(16)



3 Related Works



Table2C-

JCARprop

osalssum

mary

Work

Link

layera

ndchannelassignm

ent

Traffi

camprouting

Topo

logy

Objectiv

eampfairn

ess

Noteslimitatio

nsCA

type

Interfe

rence

Techno

logy

Traffi

ctype

Traffi

cload

Routingscheme

Mapping

(1-to-119909)119909

Con

nectivity

Objectiv

eFairn

ess

[19]

SCA

B-IRM

80211

P2P

Yes

SPM

P1

Yes

Maxim

izec

ross-sectio

ngood

-put

No

Optim

ality

isno


kcapacityestim

ationin

aheavy

traffi

cloadLess

channeld

iversitydu

etolin

kmapping

[56]

DCA

SC

A-LS

B-IRM

mdashP2

PYes

MP

01

Yes

Maxim

ize120582

(scalin

gfactor)

Min-120582

Out-of-o

rder

prob

lem

andconfi

guratio

ncomplexity

onroutingSw

itching

overhead

issues

[20]

SCA-

LSB-IRM

mdashInternet

Yes

MP

Any

ForM

Rswtr

affic

Maxim

ize120582

(scalin

gfactor)

Min-120582

Radioconstraintsa

reno

tcon

sidered

intheL

POut-of-o

rder

prob

lem

duetomultip

athrouting

[36]

DCA

B-IRM

mdashP2

Psrc-desM

RsMP

Any

No

Maxim

izea

ggregated

throug

hput

No

Out-of-o

rder

prob

lem

andconfi

guratio

ncomplexity

onroutingSw

itching

overhead

issues

Optim

ality

issue

since

onlysufficientcon

ditio

nsarec

onsid

ered

[57]

SCA

B-IRM

80211

P2PInternet

Yes

SPAny

ForM

Rswtr

affic

Maxim

izem

insparelink

capacity

Min-120582

onlin

kutilizatio

n

Usin

gsufficientcon

ditio

nin

MILPmay

notlead

toglob

alop

timalsolutio

nSing

lechannelinitia

ltopo

logy

may

lead

toinfeasiblesolutio

nin

heavy

traffi

cloadscenarios

[30]

SCA-

LSTH

MInternet

Yes

MPpredefined

01

ForM

Rswtr

affic

Maxim

ize120582

(scalin

gfactor)

Min-120582

Con

sideringon

lynecessarycond

ition

may

lead

toinfeasiblesolutio

nOut-of-o

rder

prob

lem

dueto

multip

athrouting

[70]

SCA

PHM

(wo

accumulation)

80211

Internet

No

MP(M

-GW)

L25forw

arding

ge1

Yes

Minim

izem

axlin

kutilizatio

nNo

Radioor

interfe

rencec

onstr

aintsa

reno

taccoun

tedin

initialtraffi

cloadestim

ation

Assum

esam

etrafficd

emands

inallm

eshrouters

Out-of-o

rder

prob

lem

duetomultip

athrouting

Link

rateassig

nmentisc

onsid

ered

[31]

DCA

PRMIRM

Internet

Yes

MP(M

-GW)

Any

ForM

Rswtr

affic

Maxim

izea

ggregated

throug

hput

Min-120582

Out-of-o

rder

prob

lem

andconfi

guratio

ncomplexity

onroutingEa

chradiocanbe

operated

onas

ubseto

fchann

elsC

ombinatio

nof

multip

lechannelsto

singlec

hann

elispo

ssible

[32]

SCA

THM

Internet

Yes

SPAny

Yes

Metric

onho

pcoun

tampinterfe

rence

No

Solvingsubp

roblem

with

localcon

straintsdo

esno

tguarantee

glob

alop


received

power

atan

odew

henanothern

odeis

transm

ittingisassumed

available

[62]

DCA

IRMP

RM

PHM

P2P

No

SPM

PAny

No

Maxim

izea

ggregated

throug

hput

Max-to

-Min


atmessrou

tersIn

single-path

routingop

timality

isno

tguaranteed

since

mathematicalform

ulationisform

ultip

ath

[61]

DCA

B-IRM

P2P

Yes

MP

Any

No

Maxim

ize120582

(scalin

gfactor)

Min-120582

Theq

ualityof

thes

olutionishigh

lydepend

ento

nthep

erform

ance

ofcriticalM

ISfin

ding

algorithm

[33]

SCA

B-IRM

80211

Internet

Lower

ampup

per

boun

dMP(M

GW)

Any

ForM

Rswtr

affic

Maxim

izea

ggregated

throug

hput

Boun

dedfairn

ess

Develo

pedforp

lann

ingsta

gewith

different

resourcesc

onstr

aintsNot

practic

alforlarge

networks

since

fitnessfunctio

ninvolves

solving

entiren

etworkLP

[29]

SCA-

LSTH

MP2

PYes

MP(predefin

ed)

1Yes

Maxim

izea

ggrachieved

demands

No

Link

loadsu

sedto

theC

Aaree

stimated

with

out

consideringradioor

interfe

rencec

onstr

aintsLess

channeld

iversitydu

etolin

kmapping


Table2Con

tinued

Work

Link

layera

ndchannelassignm

ent

Traffi

camprouting

Topo

logy

Objectiv

eampfairn

ess

Noteslimitatio

nsCA

type

Interfe

rence

Techno

logy

Traffi

ctype

Traffi

cload

Routingscheme

Mapping

(1-to-119909)119909

Con

nectivity

Objectiv

eFairn

ess

[37]

SCA

EMInternet

Num

bero

factiv

eflow

sMP(SPper

follo

w)

1Yes

Minim

izec

ontention

(CA)

No

Link

loadse

stimated

with

outcon

sideringradioor

interfe

rencec

onstr

aintsNot

clearlysta

tedwhere

andho

wto

implem

entfl

owratecontrol

[63]

SCA

B-IRM

80211

Internet

Yessequ

ence

oftraffi

cMP

Any

ForM

Rswtr

affic

Minim

izethe

maxim

umcongestio

n(ofw

orst

case)

No

Num

bero

ffacetsincreasee

xpon

entia

llyin

large-scalen


then

umbero

fconstraintsinLP

prob

lemhow

everrelaxingthe

convex

hullto

boxrang

esaffectsthee

fficiency

NoteSCA

static

channelassignm

entDCA

dyn

amicchannelassignm

entLS

link

schedu

lingIM

interfe

rencem

odel

IRMin

terfe

rence-rangem

odel

PRMprotocolm

odel

PHMphysic

almod

elB-bidire

ctional

THMtwo-ho

pmod

elEM

empiric

almeasurementP2

Ppeer-to

-peerSP

single-pathM

Pmultip

athMGWm

ultig

atew

ayM

Rmeshrouterand

LLlinklayer



31 C-JCAR Proposals








Table3State-aw

arec

hann

elassig

nmentrou

tingprop

osalssum

mary

Work

Statea

ware

Approach

Noteslimitatio

nsCh

annelreassignm

ent

Flow

rerouting

[35]

Costisb

ased

onchannel-link

reassig

nment

Not

applicablesin

ceafullyconn

ected

topo

logy

andfix

edpredefinedpaths

arec

onsid

ered

(i)Re

orderc

hann

elin

then

ewCA

toredu

cethec

ost

durin

gmigratio

nfro

mexistingclu

stertothen

ewclu

ster

(i)Out-of-o

rder

prob

lem

(ii)A

ssigning

channelsto

alllogicallin

ksaffectsthe

channeld

iversity

(iii)Perfo

rmance


sequ

encesa

rehigh

lyun

correlated

[85]

Costisb

ased

onchannel-link

reassig

nment

Not

applicablesin

ceafullyconn

ected

topo

logy

isconsidered

(i)Previous

channelsarer

eassignedto

links

ifthey

area

mon

gthes

etof

channelswhich

bestredu

cethe

interfe

rencea

nddo


(i)Internettraffi


atthe

gateway

(ii)P

erform

ance

degrades

with

smalln

umbero

fradios

duetorip

plee

ffecton

channelreassignm

ent

(iii)Assigning

channelsto

alllogicallin

ksaffects

channeld

iversity

[29]

Costisb

ased

onchannel-link

reassig

nment

Not

applicablesin

ceafullyconn

ected

topo

logy

andpredefinedpathsa

reconsidered

(i)Eff

ectiv

edatafor

each

channelisc

ompu

tedand

channelw

iththeh

ighesteffectiv

ecapacity

isselected

Ifthec

hann

elisno

tpreviou

slyassig

nedto

thatlin

kreconfi

guratio

ncostissubtracted

from

thee

ffective

datavalue

(i)Out-of-o

rder

prob

lem

duetomultip

ath

(ii)A

ssigning

channelsto

alllogicallin

ksaffects

channeld

iversity

(iii)Th

echann

elassig

nedto

alinkwill

beconsidered

asfin

alassig

nmentthusearliera

ssignm

entsaren

otaw

areo

flater

assig

nments

thismay

lead

toan

improp

erfin

alchannelassignm

ent

(iv)D

ueto

radioconstraint

somelinks

will

beforced

toassig

nmorec

ongeste


result

inineffi

cientchann

elassig

nment

[84]

Costisb

ased

onchannel-n

ode

reassig

nment

Not

addressed

(i)Startsfro

mthee

xisting

CAandfin

dsnewCA

bymakingsomea

djustm

ento

nthep

reviou

sone

(ii)N

umbero

fradio-chann

elreassig

nmentsis

constrainedto

apredefin

ednu

mber

(i)Perfo

rmance

degrades

with

smalln

umbero

fradios

duetorip

plee

ffecton

channelreassignm

ent

(ii)T


solved

with

CAlin

ksto

optim

izethe

networkperfo

rmance

[37]

Costisb

ased

onchannel-link

reassig

nment

Not

addressed

(i)Startin

gfro


entan

ewchannelisa

ssignedto

alinkifitdo

esno

tviolatethe

radioconstraintsa

ndredu

cesthe

interfe

rence

(ii)N

umbero


entsis

constrainedto

apredefin

ednu

mber

(i)Perfo

rmance

degrades

with

smalln

umbero

fradio

interfa

cesa

nddu

etothatreassig

nmentisrestricted

tochannelsthatdo

notvioletthe

radioconstraints

(ii)A

ssigning

channelsto

alllogicallin

ksaffects

channeld

iversity

(iii)Flow

-levelrou

tingintro

duceso

verheadto

the

routingprocedurea

ndrequ

iresa

routingentryof

each

flowin

routingtablesTh

isisim

practic

alfor

high

numbero

fflow

s

[58]

Not

addressed

Con

sideredw

here

reroutingcostis

basedon

then

umbero

fflow

shaving

different

routingpath

(i)Heuris

ticroutingbasedon

Dijk

straa

lgorith

m(ii)Th

enum

bero

fflow

swhich

allowed

having

new

routes

isconstrainedto

apredefin

edvalue

(i)Flow

reroutingcostisno

taccurately

defin

edFor

insta

ncetwonewflo

wpath

assig

nmentsmay

result

indifferent

disrup

tionwhilebo

tharec

onsid

ered

equally

inreroutingcost

























































Notations










(119906)ℎ119894




1when 119897

2is

active

Competing Interests


References
















































































































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation











n

m

z

y

u

x

| minus x||nminusx|

|uminusx|



(a)




n

m

y

z

u

x

u y z n m

(b)



Interference range

z

n

m

y

u

x














Directional link


Radio interface




64]





0119891ge 120582 lowast 120596



120583 ge0max0min

(5)



maximize sum

119891isinFLlog (0

119891) (6)















Link capacity (119897) =119905119897

119879sdot 119862119897 (7)






VC (119897) =119892 (119897)

sum1198971015840isin119868(119897)cup119897

119892 (1198971015840)sdot 119862119874 119897 isin 119871 (8)


VC (119897) + sum

1198971015840isin119868(119897)

VC (1198971015840) = 119862119874

forall1198971015840isin 119871 (9)


119892 (119897)

VC (119897)le 1 forall119897 isin 119871 (10)



sum

1198971015840isin119868(119897)cup119897

119892 (1198971015840)

1198621198971015840

le 120579 forall119897 isin 119871 (11)




sum

1198971015840isin119868(119897)cup119897

119892 (1198971015840)

1198621198971015840

le 1 forall119897 isin 119871 (12)


Time slots


Wireless link

Link

s e2

e5

e1

e4e3

e2(03)

e5(04)e1(02)

e4(03)

e3(04)





Minimize sum

1198971

sum

1198972isin119868(1198971)

119892 (1198971) lowast 119892 (119897


1| 1198972) (13)




SCA


general


sum

1198971015840isin119868(119897)cup119897

119892 (1198971015840)

1198621198971015840

le 120579 forall119897 isin 119871


sum

1198971015840isin119868(119897)

119892 (1198971015840)

1198621198971015840


IPsum

119894isin119862

ℎ119894

119898le 119870 (119906) forall119906 isin 119873

LPsum

119897

ℎ(119897)=119898

or 119905(119897)=119898

119892 (119897)

119862119897

le 119870 (119906) forall119906 isin 119873


[37]


1198971015840isin119868(119897)cup119897


119892 (119897)

VC (119897)le 1 forall119897 isin 119871

DCASCA-LS


general




1198971015840isin119868(119897)cup119897



sum

1198971015840isin119868(119897)cup119897

119892 (1198971015840)

1198621198971015840

le 120579 forall119897 isin 119871


sum

1198971015840isin119868(119897)

119892 (1198971015840)

1198621198971015840


IPsum

119897isin119871

119905(119897)=119906

or ℎ(119897)=119906

119910119905(119897) le 119870(119906) forall119906 isin 119873

LPsum

119897isin119871

119905(119897)=119906

or ℎ(119897)=119906

119892 (119897)

119862119897

le 119870 (119906) forall119906 isin 119873


[56]


sum

1198971015840isin

119864(119899)cup119864(119898)



sum

1198971015840isin119868(119897)

119892 (1198971015840)

1198621198971015840




119905(1198971015840

) le 1 forall1198971015840

isin 119868 (119897)


sum

1015840

isin119868(119897)

119892 (1198971015840

)

1198621015840


IPsum

isin119871

119905(119897)=119906

or ℎ(119897)=119906

119910119905(119897) le 119870 (119906) forall119906 isin 119873

LP

sum

isin

119905(119897)=119903

or ℎ(119897)=119903

119892 (119897)

119862




sum

1015840

isinIR(119897)cup119897

119892 (1198971015840

)

1198621015840





Connectivity

Link mapping

Topology info

Channelassignment

Interferencemodel

Technology

Traffic type

Traffic profile

Traffic collection

Routing scheme

Routing table

Execution (Re-)



Algorithms
















Topology

Approaches Traffic

Routing layer

Link layer





1when



119895 Let 119860 isin 119860

119895 Δ(119860) be the set


sum

119890isinΔ(119860)

119909 (119897) le 120593 forall119895 119860 isin 119860119895 (14)






1198821198972

1198971

= 120573 lowast119875V (119903)

119875V (119906)

forall1198971 1198972isin 119871 amp 119897

1= (119903 119904 119888) 1198972 = (119906 V 119888)

(15)



119892 (1198971) + sum

1198972isin1198711198971ampCh(119897

2)=Ch(119897

1)

119873(1198821198972

1198971

) sdot 119892 (1198972) le 120599 (119897

1)


(16)



3 Related Works



Table2C-

JCARprop

osalssum

mary

Work

Link

layera

ndchannelassignm

ent

Traffi

camprouting

Topo

logy

Objectiv

eampfairn

ess

Noteslimitatio

nsCA

type

Interfe

rence

Techno

logy

Traffi

ctype

Traffi

cload

Routingscheme

Mapping

(1-to-119909)119909

Con

nectivity

Objectiv

eFairn

ess

[19]

SCA

B-IRM

80211

P2P

Yes

SPM

P1

Yes

Maxim

izec

ross-sectio

ngood

-put

No

Optim

ality

isno


kcapacityestim

ationin

aheavy

traffi

cloadLess

channeld

iversitydu

etolin

kmapping

[56]

DCA

SC

A-LS

B-IRM

mdashP2

PYes

MP

01

Yes

Maxim

ize120582

(scalin

gfactor)

Min-120582

Out-of-o

rder

prob

lem

andconfi

guratio

ncomplexity

onroutingSw

itching

overhead

issues

[20]

SCA-

LSB-IRM

mdashInternet

Yes

MP

Any

ForM

Rswtr

affic

Maxim

ize120582

(scalin

gfactor)

Min-120582

Radioconstraintsa

reno

tcon

sidered

intheL

POut-of-o

rder

prob

lem

duetomultip

athrouting

[36]

DCA

B-IRM

mdashP2

Psrc-desM

RsMP

Any

No

Maxim

izea

ggregated

throug

hput

No

Out-of-o

rder

prob

lem

andconfi

guratio

ncomplexity

onroutingSw

itching

overhead

issues

Optim

ality

issue

since

onlysufficientcon

ditio

nsarec

onsid

ered

[57]

SCA

B-IRM

80211

P2PInternet

Yes

SPAny

ForM

Rswtr

affic

Maxim

izem

insparelink

capacity

Min-120582

onlin

kutilizatio

n

Usin

gsufficientcon

ditio

nin

MILPmay

notlead

toglob

alop

timalsolutio

nSing

lechannelinitia

ltopo

logy

may

lead

toinfeasiblesolutio

nin

heavy

traffi

cloadscenarios

[30]

SCA-

LSTH

MInternet

Yes

MPpredefined

01

ForM

Rswtr

affic

Maxim

ize120582

(scalin

gfactor)

Min-120582

Con

sideringon

lynecessarycond

ition

may

lead

toinfeasiblesolutio

nOut-of-o

rder

prob

lem

dueto

multip

athrouting

[70]

SCA

PHM

(wo

accumulation)

80211

Internet

No

MP(M

-GW)

L25forw

arding

ge1

Yes

Minim

izem

axlin

kutilizatio

nNo

Radioor

interfe

rencec

onstr

aintsa

reno

taccoun

tedin

initialtraffi

cloadestim

ation

Assum

esam

etrafficd

emands

inallm

eshrouters

Out-of-o

rder

prob

lem

duetomultip

athrouting

Link

rateassig

nmentisc

onsid

ered

[31]

DCA

PRMIRM

Internet

Yes

MP(M

-GW)

Any

ForM

Rswtr

affic

Maxim

izea

ggregated

throug

hput

Min-120582

Out-of-o

rder

prob

lem

andconfi

guratio

ncomplexity

onroutingEa

chradiocanbe

operated

onas

ubseto

fchann

elsC

ombinatio

nof

multip

lechannelsto

singlec

hann

elispo

ssible

[32]

SCA

THM

Internet

Yes

SPAny

Yes

Metric

onho

pcoun

tampinterfe

rence

No

Solvingsubp

roblem

with

localcon

straintsdo

esno

tguarantee

glob

alop


received

power

atan

odew

henanothern

odeis

transm

ittingisassumed

available

[62]

DCA

IRMP

RM

PHM

P2P

No

SPM

PAny

No

Maxim

izea

ggregated

throug

hput

Max-to

-Min


atmessrou

tersIn

single-path

routingop

timality

isno

tguaranteed

since

mathematicalform

ulationisform

ultip

ath

[61]

DCA

B-IRM

P2P

Yes

MP

Any

No

Maxim

ize120582

(scalin

gfactor)

Min-120582

Theq

ualityof

thes

olutionishigh

lydepend

ento

nthep

erform

ance

ofcriticalM

ISfin

ding

algorithm

[33]

SCA

B-IRM

80211

Internet

Lower

ampup

per

boun

dMP(M

GW)

Any

ForM

Rswtr

affic

Maxim

izea

ggregated

throug

hput

Boun

dedfairn

ess

Develo

pedforp

lann

ingsta

gewith

different

resourcesc

onstr

aintsNot

practic

alforlarge

networks

since

fitnessfunctio

ninvolves

solving

entiren

etworkLP

[29]

SCA-

LSTH

MP2

PYes

MP(predefin

ed)

1Yes

Maxim

izea

ggrachieved

demands

No

Link

loadsu

sedto

theC

Aaree

stimated

with

out

consideringradioor

interfe

rencec

onstr

aintsLess

channeld

iversitydu

etolin

kmapping


Table2Con

tinued

Work

Link

layera

ndchannelassignm

ent

Traffi

camprouting

Topo

logy

Objectiv

eampfairn

ess

Noteslimitatio

nsCA

type

Interfe

rence

Techno

logy

Traffi

ctype

Traffi

cload

Routingscheme

Mapping

(1-to-119909)119909

Con

nectivity

Objectiv

eFairn

ess

[37]

SCA

EMInternet

Num

bero

factiv

eflow

sMP(SPper

follo

w)

1Yes

Minim

izec

ontention

(CA)

No

Link

loadse

stimated

with

outcon

sideringradioor

interfe

rencec

onstr

aintsNot

clearlysta

tedwhere

andho

wto

implem

entfl

owratecontrol

[63]

SCA

B-IRM

80211

Internet

Yessequ

ence

oftraffi

cMP

Any

ForM

Rswtr

affic

Minim

izethe

maxim

umcongestio

n(ofw

orst

case)

No

Num

bero

ffacetsincreasee

xpon

entia

llyin

large-scalen


then

umbero

fconstraintsinLP

prob

lemhow

everrelaxingthe

convex

hullto

boxrang

esaffectsthee

fficiency

NoteSCA

static

channelassignm

entDCA

dyn

amicchannelassignm

entLS

link

schedu

lingIM

interfe

rencem

odel

IRMin

terfe

rence-rangem

odel

PRMprotocolm

odel

PHMphysic

almod

elB-bidire

ctional

THMtwo-ho

pmod

elEM

empiric

almeasurementP2

Ppeer-to

-peerSP

single-pathM

Pmultip

athMGWm

ultig

atew

ayM

Rmeshrouterand

LLlinklayer



31 C-JCAR Proposals








Table3State-aw

arec

hann

elassig

nmentrou

tingprop

osalssum

mary

Work

Statea

ware

Approach

Noteslimitatio

nsCh

annelreassignm

ent

Flow

rerouting

[35]

Costisb

ased

onchannel-link

reassig

nment

Not

applicablesin

ceafullyconn

ected

topo

logy

andfix

edpredefinedpaths

arec

onsid

ered

(i)Re

orderc

hann

elin

then

ewCA

toredu

cethec

ost

durin

gmigratio

nfro

mexistingclu

stertothen

ewclu

ster

(i)Out-of-o

rder

prob

lem

(ii)A

ssigning

channelsto

alllogicallin

ksaffectsthe

channeld

iversity

(iii)Perfo

rmance


sequ

encesa

rehigh

lyun

correlated

[85]

Costisb

ased

onchannel-link

reassig

nment

Not

applicablesin

ceafullyconn

ected

topo

logy

isconsidered

(i)Previous

channelsarer

eassignedto

links

ifthey

area

mon

gthes

etof

channelswhich

bestredu

cethe

interfe

rencea

nddo


(i)Internettraffi


atthe

gateway

(ii)P

erform

ance

degrades

with

smalln

umbero

fradios

duetorip

plee

ffecton

channelreassignm

ent

(iii)Assigning

channelsto

alllogicallin

ksaffects

channeld

iversity

[29]

Costisb

ased

onchannel-link

reassig

nment

Not

applicablesin

ceafullyconn

ected

topo

logy

andpredefinedpathsa

reconsidered

(i)Eff

ectiv

edatafor

each

channelisc

ompu

tedand

channelw

iththeh

ighesteffectiv

ecapacity

isselected

Ifthec

hann

elisno

tpreviou

slyassig

nedto

thatlin

kreconfi

guratio

ncostissubtracted

from

thee

ffective

datavalue

(i)Out-of-o

rder

prob

lem

duetomultip

ath

(ii)A

ssigning

channelsto

alllogicallin

ksaffects

channeld

iversity

(iii)Th

echann

elassig

nedto

alinkwill

beconsidered

asfin

alassig

nmentthusearliera

ssignm

entsaren

otaw

areo

flater

assig

nments

thismay

lead

toan

improp

erfin

alchannelassignm

ent

(iv)D

ueto

radioconstraint

somelinks

will

beforced

toassig

nmorec

ongeste


result

inineffi

cientchann

elassig

nment

[84]

Costisb

ased

onchannel-n

ode

reassig

nment

Not

addressed

(i)Startsfro

mthee

xisting

CAandfin

dsnewCA

bymakingsomea

djustm

ento

nthep

reviou

sone

(ii)N

umbero

fradio-chann

elreassig

nmentsis

constrainedto

apredefin

ednu

mber

(i)Perfo

rmance

degrades

with

smalln

umbero

fradios

duetorip

plee

ffecton

channelreassignm

ent

(ii)T


solved

with

CAlin

ksto

optim

izethe

networkperfo

rmance

[37]

Costisb

ased

onchannel-link

reassig

nment

Not

addressed

(i)Startin

gfro


entan

ewchannelisa

ssignedto

alinkifitdo

esno

tviolatethe

radioconstraintsa

ndredu

cesthe

interfe

rence

(ii)N

umbero


entsis

constrainedto

apredefin

ednu

mber

(i)Perfo

rmance

degrades

with

smalln

umbero

fradio

interfa

cesa

nddu

etothatreassig

nmentisrestricted

tochannelsthatdo

notvioletthe

radioconstraints

(ii)A

ssigning

channelsto

alllogicallin

ksaffects

channeld

iversity

(iii)Flow

-levelrou

tingintro

duceso

verheadto

the

routingprocedurea

ndrequ

iresa

routingentryof

each

flowin

routingtablesTh

isisim

practic

alfor

high

numbero

fflow

s

[58]

Not

addressed

Con

sideredw

here

reroutingcostis

basedon

then

umbero

fflow

shaving

different

routingpath

(i)Heuris

ticroutingbasedon

Dijk

straa

lgorith

m(ii)Th

enum

bero

fflow

swhich

allowed

having

new

routes

isconstrainedto

apredefin

edvalue

(i)Flow

reroutingcostisno

taccurately

defin

edFor

insta

ncetwonewflo

wpath

assig

nmentsmay

result

indifferent

disrup

tionwhilebo

tharec

onsid

ered

equally

inreroutingcost

























































Notations










(119906)ℎ119894




1when 119897

2is

active

Competing Interests


References
















































































































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation












Directional link


Radio interface




64]





0119891ge 120582 lowast 120596



120583 ge0max0min

(5)



maximize sum

119891isinFLlog (0

119891) (6)















Link capacity (119897) =119905119897

119879sdot 119862119897 (7)






VC (119897) =119892 (119897)

sum1198971015840isin119868(119897)cup119897

119892 (1198971015840)sdot 119862119874 119897 isin 119871 (8)


VC (119897) + sum

1198971015840isin119868(119897)

VC (1198971015840) = 119862119874

forall1198971015840isin 119871 (9)


119892 (119897)

VC (119897)le 1 forall119897 isin 119871 (10)



sum

1198971015840isin119868(119897)cup119897

119892 (1198971015840)

1198621198971015840

le 120579 forall119897 isin 119871 (11)




sum

1198971015840isin119868(119897)cup119897

119892 (1198971015840)

1198621198971015840

le 1 forall119897 isin 119871 (12)


Time slots


Wireless link

Link

s e2

e5

e1

e4e3

e2(03)

e5(04)e1(02)

e4(03)

e3(04)





Minimize sum

1198971

sum

1198972isin119868(1198971)

119892 (1198971) lowast 119892 (119897


1| 1198972) (13)




SCA


general


sum

1198971015840isin119868(119897)cup119897

119892 (1198971015840)

1198621198971015840

le 120579 forall119897 isin 119871


sum

1198971015840isin119868(119897)

119892 (1198971015840)

1198621198971015840


IPsum

119894isin119862

ℎ119894

119898le 119870 (119906) forall119906 isin 119873

LPsum

119897

ℎ(119897)=119898

or 119905(119897)=119898

119892 (119897)

119862119897

le 119870 (119906) forall119906 isin 119873


[37]


1198971015840isin119868(119897)cup119897


119892 (119897)

VC (119897)le 1 forall119897 isin 119871

DCASCA-LS


general




1198971015840isin119868(119897)cup119897



sum

1198971015840isin119868(119897)cup119897

119892 (1198971015840)

1198621198971015840

le 120579 forall119897 isin 119871


sum

1198971015840isin119868(119897)

119892 (1198971015840)

1198621198971015840


IPsum

119897isin119871

119905(119897)=119906

or ℎ(119897)=119906

119910119905(119897) le 119870(119906) forall119906 isin 119873

LPsum

119897isin119871

119905(119897)=119906

or ℎ(119897)=119906

119892 (119897)

119862119897

le 119870 (119906) forall119906 isin 119873


[56]


sum

1198971015840isin

119864(119899)cup119864(119898)



sum

1198971015840isin119868(119897)

119892 (1198971015840)

1198621198971015840




119905(1198971015840

) le 1 forall1198971015840

isin 119868 (119897)


sum

1015840

isin119868(119897)

119892 (1198971015840

)

1198621015840


IPsum

isin119871

119905(119897)=119906

or ℎ(119897)=119906

119910119905(119897) le 119870 (119906) forall119906 isin 119873

LP

sum

isin

119905(119897)=119903

or ℎ(119897)=119903

119892 (119897)

119862




sum

1015840

isinIR(119897)cup119897

119892 (1198971015840

)

1198621015840





Connectivity

Link mapping

Topology info

Channelassignment

Interferencemodel

Technology

Traffic type

Traffic profile

Traffic collection

Routing scheme

Routing table

Execution (Re-)



Algorithms
















Topology

Approaches Traffic

Routing layer

Link layer





1when



119895 Let 119860 isin 119860

119895 Δ(119860) be the set


sum

119890isinΔ(119860)

119909 (119897) le 120593 forall119895 119860 isin 119860119895 (14)






1198821198972

1198971

= 120573 lowast119875V (119903)

119875V (119906)

forall1198971 1198972isin 119871 amp 119897

1= (119903 119904 119888) 1198972 = (119906 V 119888)

(15)



119892 (1198971) + sum

1198972isin1198711198971ampCh(119897

2)=Ch(119897

1)

119873(1198821198972

1198971

) sdot 119892 (1198972) le 120599 (119897

1)


(16)



3 Related Works



Table2C-

JCARprop

osalssum

mary

Work

Link

layera

ndchannelassignm

ent

Traffi

camprouting

Topo

logy

Objectiv

eampfairn

ess

Noteslimitatio

nsCA

type

Interfe

rence

Techno

logy

Traffi

ctype

Traffi

cload

Routingscheme

Mapping

(1-to-119909)119909

Con

nectivity

Objectiv

eFairn

ess

[19]

SCA

B-IRM

80211

P2P

Yes

SPM

P1

Yes

Maxim

izec

ross-sectio

ngood

-put

No

Optim

ality

isno


kcapacityestim

ationin

aheavy

traffi

cloadLess

channeld

iversitydu

etolin

kmapping

[56]

DCA

SC

A-LS

B-IRM

mdashP2

PYes

MP

01

Yes

Maxim

ize120582

(scalin

gfactor)

Min-120582

Out-of-o

rder

prob

lem

andconfi

guratio

ncomplexity

onroutingSw

itching

overhead

issues

[20]

SCA-

LSB-IRM

mdashInternet

Yes

MP

Any

ForM

Rswtr

affic

Maxim

ize120582

(scalin

gfactor)

Min-120582

Radioconstraintsa

reno

tcon

sidered

intheL

POut-of-o

rder

prob

lem

duetomultip

athrouting

[36]

DCA

B-IRM

mdashP2

Psrc-desM

RsMP

Any

No

Maxim

izea

ggregated

throug

hput

No

Out-of-o

rder

prob

lem

andconfi

guratio

ncomplexity

onroutingSw

itching

overhead

issues

Optim

ality

issue

since

onlysufficientcon

ditio

nsarec

onsid

ered

[57]

SCA

B-IRM

80211

P2PInternet

Yes

SPAny

ForM

Rswtr

affic

Maxim

izem

insparelink

capacity

Min-120582

onlin

kutilizatio

n

Usin

gsufficientcon

ditio

nin

MILPmay

notlead

toglob

alop

timalsolutio

nSing

lechannelinitia

ltopo

logy

may

lead

toinfeasiblesolutio

nin

heavy

traffi

cloadscenarios

[30]

SCA-

LSTH

MInternet

Yes

MPpredefined

01

ForM

Rswtr

affic

Maxim

ize120582

(scalin

gfactor)

Min-120582

Con

sideringon

lynecessarycond

ition

may

lead

toinfeasiblesolutio

nOut-of-o

rder

prob

lem

dueto

multip

athrouting

[70]

SCA

PHM

(wo

accumulation)

80211

Internet

No

MP(M

-GW)

L25forw

arding

ge1

Yes

Minim

izem

axlin

kutilizatio

nNo

Radioor

interfe

rencec

onstr

aintsa

reno

taccoun

tedin

initialtraffi

cloadestim

ation

Assum

esam

etrafficd

emands

inallm

eshrouters

Out-of-o

rder

prob

lem

duetomultip

athrouting

Link

rateassig

nmentisc

onsid

ered

[31]

DCA

PRMIRM

Internet

Yes

MP(M

-GW)

Any

ForM

Rswtr

affic

Maxim

izea

ggregated

throug

hput

Min-120582

Out-of-o

rder

prob

lem

andconfi

guratio

ncomplexity

onroutingEa

chradiocanbe

operated

onas

ubseto

fchann

elsC

ombinatio

nof

multip

lechannelsto

singlec

hann

elispo

ssible

[32]

SCA

THM

Internet

Yes

SPAny

Yes

Metric

onho

pcoun

tampinterfe

rence

No

Solvingsubp

roblem

with

localcon

straintsdo

esno

tguarantee

glob

alop


received

power

atan

odew

henanothern

odeis

transm

ittingisassumed

available

[62]

DCA

IRMP

RM

PHM

P2P

No

SPM

PAny

No

Maxim

izea

ggregated

throug

hput

Max-to

-Min


atmessrou

tersIn

single-path

routingop

timality

isno

tguaranteed

since

mathematicalform

ulationisform

ultip

ath

[61]

DCA

B-IRM

P2P

Yes

MP

Any

No

Maxim

ize120582

(scalin

gfactor)

Min-120582

Theq

ualityof

thes

olutionishigh

lydepend

ento

nthep

erform

ance

ofcriticalM

ISfin

ding

algorithm

[33]

SCA

B-IRM

80211

Internet

Lower

ampup

per

boun

dMP(M

GW)

Any

ForM

Rswtr

affic

Maxim

izea

ggregated

throug

hput

Boun

dedfairn

ess

Develo

pedforp

lann

ingsta

gewith

different

resourcesc

onstr

aintsNot

practic

alforlarge

networks

since

fitnessfunctio

ninvolves

solving

entiren

etworkLP

[29]

SCA-

LSTH

MP2

PYes

MP(predefin

ed)

1Yes

Maxim

izea

ggrachieved

demands

No

Link

loadsu

sedto

theC

Aaree

stimated

with

out

consideringradioor

interfe

rencec

onstr

aintsLess

channeld

iversitydu

etolin

kmapping


Table2Con

tinued

Work

Link

layera

ndchannelassignm

ent

Traffi

camprouting

Topo

logy

Objectiv

eampfairn

ess

Noteslimitatio

nsCA

type

Interfe

rence

Techno

logy

Traffi

ctype

Traffi

cload

Routingscheme

Mapping

(1-to-119909)119909

Con

nectivity

Objectiv

eFairn

ess

[37]

SCA

EMInternet

Num

bero

factiv

eflow

sMP(SPper

follo

w)

1Yes

Minim

izec

ontention

(CA)

No

Link

loadse

stimated

with

outcon

sideringradioor

interfe

rencec

onstr

aintsNot

clearlysta

tedwhere

andho

wto

implem

entfl

owratecontrol

[63]

SCA

B-IRM

80211

Internet

Yessequ

ence

oftraffi

cMP

Any

ForM

Rswtr

affic

Minim

izethe

maxim

umcongestio

n(ofw

orst

case)

No

Num

bero

ffacetsincreasee

xpon

entia

llyin

large-scalen


then

umbero

fconstraintsinLP

prob

lemhow

everrelaxingthe

convex

hullto

boxrang

esaffectsthee

fficiency

NoteSCA

static

channelassignm

entDCA

dyn

amicchannelassignm

entLS

link

schedu

lingIM

interfe

rencem

odel

IRMin

terfe

rence-rangem

odel

PRMprotocolm

odel

PHMphysic

almod

elB-bidire

ctional

THMtwo-ho

pmod

elEM

empiric

almeasurementP2

Ppeer-to

-peerSP

single-pathM

Pmultip

athMGWm

ultig

atew

ayM

Rmeshrouterand

LLlinklayer



31 C-JCAR Proposals








Table3State-aw

arec

hann

elassig

nmentrou

tingprop

osalssum

mary

Work

Statea

ware

Approach

Noteslimitatio

nsCh

annelreassignm

ent

Flow

rerouting

[35]

Costisb

ased

onchannel-link

reassig

nment

Not

applicablesin

ceafullyconn

ected

topo

logy

andfix

edpredefinedpaths

arec

onsid

ered

(i)Re

orderc

hann

elin

then

ewCA

toredu

cethec

ost

durin

gmigratio

nfro

mexistingclu

stertothen

ewclu

ster

(i)Out-of-o

rder

prob

lem

(ii)A

ssigning

channelsto

alllogicallin

ksaffectsthe

channeld

iversity

(iii)Perfo

rmance


sequ

encesa

rehigh

lyun

correlated

[85]

Costisb

ased

onchannel-link

reassig

nment

Not

applicablesin

ceafullyconn

ected

topo

logy

isconsidered

(i)Previous

channelsarer

eassignedto

links

ifthey

area

mon

gthes

etof

channelswhich

bestredu

cethe

interfe

rencea

nddo


(i)Internettraffi


atthe

gateway

(ii)P

erform

ance

degrades

with

smalln

umbero

fradios

duetorip

plee

ffecton

channelreassignm

ent

(iii)Assigning

channelsto

alllogicallin

ksaffects

channeld

iversity

[29]

Costisb

ased

onchannel-link

reassig

nment

Not

applicablesin

ceafullyconn

ected

topo

logy

andpredefinedpathsa

reconsidered

(i)Eff

ectiv

edatafor

each

channelisc

ompu

tedand

channelw

iththeh

ighesteffectiv

ecapacity

isselected

Ifthec

hann

elisno

tpreviou

slyassig

nedto

thatlin

kreconfi

guratio

ncostissubtracted

from

thee

ffective

datavalue

(i)Out-of-o

rder

prob

lem

duetomultip

ath

(ii)A

ssigning

channelsto

alllogicallin

ksaffects

channeld

iversity

(iii)Th

echann

elassig

nedto

alinkwill

beconsidered

asfin

alassig

nmentthusearliera

ssignm

entsaren

otaw

areo

flater

assig

nments

thismay

lead

toan

improp

erfin

alchannelassignm

ent

(iv)D

ueto

radioconstraint

somelinks

will

beforced

toassig

nmorec

ongeste


result

inineffi

cientchann

elassig

nment

[84]

Costisb

ased

onchannel-n

ode

reassig

nment

Not

addressed

(i)Startsfro

mthee

xisting

CAandfin

dsnewCA

bymakingsomea

djustm

ento

nthep

reviou

sone

(ii)N

umbero

fradio-chann

elreassig

nmentsis

constrainedto

apredefin

ednu

mber

(i)Perfo

rmance

degrades

with

smalln

umbero

fradios

duetorip

plee

ffecton

channelreassignm

ent

(ii)T


solved

with

CAlin

ksto

optim

izethe

networkperfo

rmance

[37]

Costisb

ased

onchannel-link

reassig

nment

Not

addressed

(i)Startin

gfro


entan

ewchannelisa

ssignedto

alinkifitdo

esno

tviolatethe

radioconstraintsa

ndredu

cesthe

interfe

rence

(ii)N

umbero


entsis

constrainedto

apredefin

ednu

mber

(i)Perfo

rmance

degrades

with

smalln

umbero

fradio

interfa

cesa

nddu

etothatreassig

nmentisrestricted

tochannelsthatdo

notvioletthe

radioconstraints

(ii)A

ssigning

channelsto

alllogicallin

ksaffects

channeld

iversity

(iii)Flow

-levelrou

tingintro

duceso

verheadto

the

routingprocedurea

ndrequ

iresa

routingentryof

each

flowin

routingtablesTh

isisim

practic

alfor

high

numbero

fflow

s

[58]

Not

addressed

Con

sideredw

here

reroutingcostis

basedon

then

umbero

fflow

shaving

different

routingpath

(i)Heuris

ticroutingbasedon

Dijk

straa

lgorith

m(ii)Th

enum

bero

fflow

swhich

allowed

having

new

routes

isconstrainedto

apredefin

edvalue

(i)Flow

reroutingcostisno

taccurately

defin

edFor

insta

ncetwonewflo

wpath

assig

nmentsmay

result

indifferent

disrup

tionwhilebo

tharec

onsid

ered

equally

inreroutingcost

























































Notations










(119906)ℎ119894




1when 119897

2is

active

Competing Interests


References
















































































































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation











Link capacity (119897) =119905119897

119879sdot 119862119897 (7)






VC (119897) =119892 (119897)

sum1198971015840isin119868(119897)cup119897

119892 (1198971015840)sdot 119862119874 119897 isin 119871 (8)


VC (119897) + sum

1198971015840isin119868(119897)

VC (1198971015840) = 119862119874

forall1198971015840isin 119871 (9)


119892 (119897)

VC (119897)le 1 forall119897 isin 119871 (10)



sum

1198971015840isin119868(119897)cup119897

119892 (1198971015840)

1198621198971015840

le 120579 forall119897 isin 119871 (11)




sum

1198971015840isin119868(119897)cup119897

119892 (1198971015840)

1198621198971015840

le 1 forall119897 isin 119871 (12)


Time slots


Wireless link

Link

s e2

e5

e1

e4e3

e2(03)

e5(04)e1(02)

e4(03)

e3(04)





Minimize sum

1198971

sum

1198972isin119868(1198971)

119892 (1198971) lowast 119892 (119897


1| 1198972) (13)




SCA


general


sum

1198971015840isin119868(119897)cup119897

119892 (1198971015840)

1198621198971015840

le 120579 forall119897 isin 119871


sum

1198971015840isin119868(119897)

119892 (1198971015840)

1198621198971015840


IPsum

119894isin119862

ℎ119894

119898le 119870 (119906) forall119906 isin 119873

LPsum

119897

ℎ(119897)=119898

or 119905(119897)=119898

119892 (119897)

119862119897

le 119870 (119906) forall119906 isin 119873


[37]


1198971015840isin119868(119897)cup119897


119892 (119897)

VC (119897)le 1 forall119897 isin 119871

DCASCA-LS


general




1198971015840isin119868(119897)cup119897



sum

1198971015840isin119868(119897)cup119897

119892 (1198971015840)

1198621198971015840

le 120579 forall119897 isin 119871


sum

1198971015840isin119868(119897)

119892 (1198971015840)

1198621198971015840


IPsum

119897isin119871

119905(119897)=119906

or ℎ(119897)=119906

119910119905(119897) le 119870(119906) forall119906 isin 119873

LPsum

119897isin119871

119905(119897)=119906

or ℎ(119897)=119906

119892 (119897)

119862119897

le 119870 (119906) forall119906 isin 119873


[56]


sum

1198971015840isin

119864(119899)cup119864(119898)



sum

1198971015840isin119868(119897)

119892 (1198971015840)

1198621198971015840




119905(1198971015840

) le 1 forall1198971015840

isin 119868 (119897)


sum

1015840

isin119868(119897)

119892 (1198971015840

)

1198621015840


IPsum

isin119871

119905(119897)=119906

or ℎ(119897)=119906

119910119905(119897) le 119870 (119906) forall119906 isin 119873

LP

sum

isin

119905(119897)=119903

or ℎ(119897)=119903

119892 (119897)

119862




sum

1015840

isinIR(119897)cup119897

119892 (1198971015840

)

1198621015840





Connectivity

Link mapping

Topology info

Channelassignment

Interferencemodel

Technology

Traffic type

Traffic profile

Traffic collection

Routing scheme

Routing table

Execution (Re-)



Algorithms
















Topology

Approaches Traffic

Routing layer

Link layer





1when



119895 Let 119860 isin 119860

119895 Δ(119860) be the set


sum

119890isinΔ(119860)

119909 (119897) le 120593 forall119895 119860 isin 119860119895 (14)






1198821198972

1198971

= 120573 lowast119875V (119903)

119875V (119906)

forall1198971 1198972isin 119871 amp 119897

1= (119903 119904 119888) 1198972 = (119906 V 119888)

(15)



119892 (1198971) + sum

1198972isin1198711198971ampCh(119897

2)=Ch(119897

1)

119873(1198821198972

1198971

) sdot 119892 (1198972) le 120599 (119897

1)


(16)



3 Related Works



Table2C-

JCARprop

osalssum

mary

Work

Link

layera

ndchannelassignm

ent

Traffi

camprouting

Topo

logy

Objectiv

eampfairn

ess

Noteslimitatio

nsCA

type

Interfe

rence

Techno

logy

Traffi

ctype

Traffi

cload

Routingscheme

Mapping

(1-to-119909)119909

Con

nectivity

Objectiv

eFairn

ess

[19]

SCA

B-IRM

80211

P2P

Yes

SPM

P1

Yes

Maxim

izec

ross-sectio

ngood

-put

No

Optim

ality

isno


kcapacityestim

ationin

aheavy

traffi

cloadLess

channeld

iversitydu

etolin

kmapping

[56]

DCA

SC

A-LS

B-IRM

mdashP2

PYes

MP

01

Yes

Maxim

ize120582

(scalin

gfactor)

Min-120582

Out-of-o

rder

prob

lem

andconfi

guratio

ncomplexity

onroutingSw

itching

overhead

issues

[20]

SCA-

LSB-IRM

mdashInternet

Yes

MP

Any

ForM

Rswtr

affic

Maxim

ize120582

(scalin

gfactor)

Min-120582

Radioconstraintsa

reno

tcon

sidered

intheL

POut-of-o

rder

prob

lem

duetomultip

athrouting

[36]

DCA

B-IRM

mdashP2

Psrc-desM

RsMP

Any

No

Maxim

izea

ggregated

throug

hput

No

Out-of-o

rder

prob

lem

andconfi

guratio

ncomplexity

onroutingSw

itching

overhead

issues

Optim

ality

issue

since

onlysufficientcon

ditio

nsarec

onsid

ered

[57]

SCA

B-IRM

80211

P2PInternet

Yes

SPAny

ForM

Rswtr

affic

Maxim

izem

insparelink

capacity

Min-120582

onlin

kutilizatio

n

Usin

gsufficientcon

ditio

nin

MILPmay

notlead

toglob

alop

timalsolutio

nSing

lechannelinitia

ltopo

logy

may

lead

toinfeasiblesolutio

nin

heavy

traffi

cloadscenarios

[30]

SCA-

LSTH

MInternet

Yes

MPpredefined

01

ForM

Rswtr

affic

Maxim

ize120582

(scalin

gfactor)

Min-120582

Con

sideringon

lynecessarycond

ition

may

lead

toinfeasiblesolutio

nOut-of-o

rder

prob

lem

dueto

multip

athrouting

[70]

SCA

PHM

(wo

accumulation)

80211

Internet

No

MP(M

-GW)

L25forw

arding

ge1

Yes

Minim

izem

axlin

kutilizatio

nNo

Radioor

interfe

rencec

onstr

aintsa

reno

taccoun

tedin

initialtraffi

cloadestim

ation

Assum

esam

etrafficd

emands

inallm

eshrouters

Out-of-o

rder

prob

lem

duetomultip

athrouting

Link

rateassig

nmentisc

onsid

ered

[31]

DCA

PRMIRM

Internet

Yes

MP(M

-GW)

Any

ForM

Rswtr

affic

Maxim

izea

ggregated

throug

hput

Min-120582

Out-of-o

rder

prob

lem

andconfi

guratio

ncomplexity

onroutingEa

chradiocanbe

operated

onas

ubseto

fchann

elsC

ombinatio

nof

multip

lechannelsto

singlec

hann

elispo

ssible

[32]

SCA

THM

Internet

Yes

SPAny

Yes

Metric

onho

pcoun

tampinterfe

rence

No

Solvingsubp

roblem

with

localcon

straintsdo

esno

tguarantee

glob

alop


received

power

atan

odew

henanothern

odeis

transm

ittingisassumed

available

[62]

DCA

IRMP

RM

PHM

P2P

No

SPM

PAny

No

Maxim

izea

ggregated

throug

hput

Max-to

-Min


atmessrou

tersIn

single-path

routingop

timality

isno

tguaranteed

since

mathematicalform

ulationisform

ultip

ath

[61]

DCA

B-IRM

P2P

Yes

MP

Any

No

Maxim

ize120582

(scalin

gfactor)

Min-120582

Theq

ualityof

thes

olutionishigh

lydepend

ento

nthep

erform

ance

ofcriticalM

ISfin

ding

algorithm

[33]

SCA

B-IRM

80211

Internet

Lower

ampup

per

boun

dMP(M

GW)

Any

ForM

Rswtr

affic

Maxim

izea

ggregated

throug

hput

Boun

dedfairn

ess

Develo

pedforp

lann

ingsta

gewith

different

resourcesc

onstr

aintsNot

practic

alforlarge

networks

since

fitnessfunctio

ninvolves

solving

entiren

etworkLP

[29]

SCA-

LSTH

MP2

PYes

MP(predefin

ed)

1Yes

Maxim

izea

ggrachieved

demands

No

Link

loadsu

sedto

theC

Aaree

stimated

with

out

consideringradioor

interfe

rencec

onstr

aintsLess

channeld

iversitydu

etolin

kmapping


Table2Con

tinued

Work

Link

layera

ndchannelassignm

ent

Traffi

camprouting

Topo

logy

Objectiv

eampfairn

ess

Noteslimitatio

nsCA

type

Interfe

rence

Techno

logy

Traffi

ctype

Traffi

cload

Routingscheme

Mapping

(1-to-119909)119909

Con

nectivity

Objectiv

eFairn

ess

[37]

SCA

EMInternet

Num

bero

factiv

eflow

sMP(SPper

follo

w)

1Yes

Minim

izec

ontention

(CA)

No

Link

loadse

stimated

with

outcon

sideringradioor

interfe

rencec

onstr

aintsNot

clearlysta

tedwhere

andho

wto

implem

entfl

owratecontrol

[63]

SCA

B-IRM

80211

Internet

Yessequ

ence

oftraffi

cMP

Any

ForM

Rswtr

affic

Minim

izethe

maxim

umcongestio

n(ofw

orst

case)

No

Num

bero

ffacetsincreasee

xpon

entia

llyin

large-scalen


then

umbero

fconstraintsinLP

prob

lemhow

everrelaxingthe

convex

hullto

boxrang

esaffectsthee

fficiency

NoteSCA

static

channelassignm

entDCA

dyn

amicchannelassignm

entLS

link

schedu

lingIM

interfe

rencem

odel

IRMin

terfe

rence-rangem

odel

PRMprotocolm

odel

PHMphysic

almod

elB-bidire

ctional

THMtwo-ho

pmod

elEM

empiric

almeasurementP2

Ppeer-to

-peerSP

single-pathM

Pmultip

athMGWm

ultig

atew

ayM

Rmeshrouterand

LLlinklayer



31 C-JCAR Proposals








Table3State-aw

arec

hann

elassig

nmentrou

tingprop

osalssum

mary

Work

Statea

ware

Approach

Noteslimitatio

nsCh

annelreassignm

ent

Flow

rerouting

[35]

Costisb

ased

onchannel-link

reassig

nment

Not

applicablesin

ceafullyconn

ected

topo

logy

andfix

edpredefinedpaths

arec

onsid

ered

(i)Re

orderc

hann

elin

then

ewCA

toredu

cethec

ost

durin

gmigratio

nfro

mexistingclu

stertothen

ewclu

ster

(i)Out-of-o

rder

prob

lem

(ii)A

ssigning

channelsto

alllogicallin

ksaffectsthe

channeld

iversity

(iii)Perfo

rmance


sequ

encesa

rehigh

lyun

correlated

[85]

Costisb

ased

onchannel-link

reassig

nment

Not

applicablesin

ceafullyconn

ected

topo

logy

isconsidered

(i)Previous

channelsarer

eassignedto

links

ifthey

area

mon

gthes

etof

channelswhich

bestredu

cethe

interfe

rencea

nddo


(i)Internettraffi


atthe

gateway

(ii)P

erform

ance

degrades

with

smalln

umbero

fradios

duetorip

plee

ffecton

channelreassignm

ent

(iii)Assigning

channelsto

alllogicallin

ksaffects

channeld

iversity

[29]

Costisb

ased

onchannel-link

reassig

nment

Not

applicablesin

ceafullyconn

ected

topo

logy

andpredefinedpathsa

reconsidered

(i)Eff

ectiv

edatafor

each

channelisc

ompu

tedand

channelw

iththeh

ighesteffectiv

ecapacity

isselected

Ifthec

hann

elisno

tpreviou

slyassig

nedto

thatlin

kreconfi

guratio

ncostissubtracted

from

thee

ffective

datavalue

(i)Out-of-o

rder

prob

lem

duetomultip

ath

(ii)A

ssigning

channelsto

alllogicallin

ksaffects

channeld

iversity

(iii)Th

echann

elassig

nedto

alinkwill

beconsidered

asfin

alassig

nmentthusearliera

ssignm

entsaren

otaw

areo

flater

assig

nments

thismay

lead

toan

improp

erfin

alchannelassignm

ent

(iv)D

ueto

radioconstraint

somelinks

will

beforced

toassig

nmorec

ongeste


result

inineffi

cientchann

elassig

nment

[84]

Costisb

ased

onchannel-n

ode

reassig

nment

Not

addressed

(i)Startsfro

mthee

xisting

CAandfin

dsnewCA

bymakingsomea

djustm

ento

nthep

reviou

sone

(ii)N

umbero

fradio-chann

elreassig

nmentsis

constrainedto

apredefin

ednu

mber

(i)Perfo

rmance

degrades

with

smalln

umbero

fradios

duetorip

plee

ffecton

channelreassignm

ent

(ii)T


solved

with

CAlin

ksto

optim

izethe

networkperfo

rmance

[37]

Costisb

ased

onchannel-link

reassig

nment

Not

addressed

(i)Startin

gfro


entan

ewchannelisa

ssignedto

alinkifitdo

esno

tviolatethe

radioconstraintsa

ndredu

cesthe

interfe

rence

(ii)N

umbero


entsis

constrainedto

apredefin

ednu

mber

(i)Perfo

rmance

degrades

with

smalln

umbero

fradio

interfa

cesa

nddu

etothatreassig

nmentisrestricted

tochannelsthatdo

notvioletthe

radioconstraints

(ii)A

ssigning

channelsto

alllogicallin

ksaffects

channeld

iversity

(iii)Flow

-levelrou

tingintro

duceso

verheadto

the

routingprocedurea

ndrequ

iresa

routingentryof

each

flowin

routingtablesTh

isisim

practic

alfor

high

numbero

fflow

s

[58]

Not

addressed

Con

sideredw

here

reroutingcostis

basedon

then

umbero

fflow

shaving

different

routingpath

(i)Heuris

ticroutingbasedon

Dijk

straa

lgorith

m(ii)Th

enum

bero

fflow

swhich

allowed

having

new

routes

isconstrainedto

apredefin

edvalue

(i)Flow

reroutingcostisno

taccurately

defin

edFor

insta

ncetwonewflo

wpath

assig

nmentsmay

result

indifferent

disrup

tionwhilebo

tharec

onsid

ered

equally

inreroutingcost

























































Notations










(119906)ℎ119894




1when 119897

2is

active

Competing Interests


References
















































































































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation












SCA


general


sum

1198971015840isin119868(119897)cup119897

119892 (1198971015840)

1198621198971015840

le 120579 forall119897 isin 119871


sum

1198971015840isin119868(119897)

119892 (1198971015840)

1198621198971015840


IPsum

119894isin119862

ℎ119894

119898le 119870 (119906) forall119906 isin 119873

LPsum

119897

ℎ(119897)=119898

or 119905(119897)=119898

119892 (119897)

119862119897

le 119870 (119906) forall119906 isin 119873


[37]


1198971015840isin119868(119897)cup119897


119892 (119897)

VC (119897)le 1 forall119897 isin 119871

DCASCA-LS


general




1198971015840isin119868(119897)cup119897



sum

1198971015840isin119868(119897)cup119897

119892 (1198971015840)

1198621198971015840

le 120579 forall119897 isin 119871


sum

1198971015840isin119868(119897)

119892 (1198971015840)

1198621198971015840


IPsum

119897isin119871

119905(119897)=119906

or ℎ(119897)=119906

119910119905(119897) le 119870(119906) forall119906 isin 119873

LPsum

119897isin119871

119905(119897)=119906

or ℎ(119897)=119906

119892 (119897)

119862119897

le 119870 (119906) forall119906 isin 119873


[56]


sum

1198971015840isin

119864(119899)cup119864(119898)



sum

1198971015840isin119868(119897)

119892 (1198971015840)

1198621198971015840




119905(1198971015840

) le 1 forall1198971015840

isin 119868 (119897)


sum

1015840

isin119868(119897)

119892 (1198971015840

)

1198621015840


IPsum

isin119871

119905(119897)=119906

or ℎ(119897)=119906

119910119905(119897) le 119870 (119906) forall119906 isin 119873

LP

sum

isin

119905(119897)=119903

or ℎ(119897)=119903

119892 (119897)

119862




sum

1015840

isinIR(119897)cup119897

119892 (1198971015840

)

1198621015840





Connectivity

Link mapping

Topology info

Channelassignment

Interferencemodel

Technology

Traffic type

Traffic profile

Traffic collection

Routing scheme

Routing table

Execution (Re-)



Algorithms
















Topology

Approaches Traffic

Routing layer

Link layer





1when



119895 Let 119860 isin 119860

119895 Δ(119860) be the set


sum

119890isinΔ(119860)

119909 (119897) le 120593 forall119895 119860 isin 119860119895 (14)






1198821198972

1198971

= 120573 lowast119875V (119903)

119875V (119906)

forall1198971 1198972isin 119871 amp 119897

1= (119903 119904 119888) 1198972 = (119906 V 119888)

(15)



119892 (1198971) + sum

1198972isin1198711198971ampCh(119897

2)=Ch(119897

1)

119873(1198821198972

1198971

) sdot 119892 (1198972) le 120599 (119897

1)


(16)



3 Related Works



Table2C-

JCARprop

osalssum

mary

Work

Link

layera

ndchannelassignm

ent

Traffi

camprouting

Topo

logy

Objectiv

eampfairn

ess

Noteslimitatio

nsCA

type

Interfe

rence

Techno

logy

Traffi

ctype

Traffi

cload

Routingscheme

Mapping

(1-to-119909)119909

Con

nectivity

Objectiv

eFairn

ess

[19]

SCA

B-IRM

80211

P2P

Yes

SPM

P1

Yes

Maxim

izec

ross-sectio

ngood

-put

No

Optim

ality

isno


kcapacityestim

ationin

aheavy

traffi

cloadLess

channeld

iversitydu

etolin

kmapping

[56]

DCA

SC

A-LS

B-IRM

mdashP2

PYes

MP

01

Yes

Maxim

ize120582

(scalin

gfactor)

Min-120582

Out-of-o

rder

prob

lem

andconfi

guratio

ncomplexity

onroutingSw

itching

overhead

issues

[20]

SCA-

LSB-IRM

mdashInternet

Yes

MP

Any

ForM

Rswtr

affic

Maxim

ize120582

(scalin

gfactor)

Min-120582

Radioconstraintsa

reno

tcon

sidered

intheL

POut-of-o

rder

prob

lem

duetomultip

athrouting

[36]

DCA

B-IRM

mdashP2

Psrc-desM

RsMP

Any

No

Maxim

izea

ggregated

throug

hput

No

Out-of-o

rder

prob

lem

andconfi

guratio

ncomplexity

onroutingSw

itching

overhead

issues

Optim

ality

issue

since

onlysufficientcon

ditio

nsarec

onsid

ered

[57]

SCA

B-IRM

80211

P2PInternet

Yes

SPAny

ForM

Rswtr

affic

Maxim

izem

insparelink

capacity

Min-120582

onlin

kutilizatio

n

Usin

gsufficientcon

ditio

nin

MILPmay

notlead

toglob

alop

timalsolutio

nSing

lechannelinitia

ltopo

logy

may

lead

toinfeasiblesolutio

nin

heavy

traffi

cloadscenarios

[30]

SCA-

LSTH

MInternet

Yes

MPpredefined

01

ForM

Rswtr

affic

Maxim

ize120582

(scalin

gfactor)

Min-120582

Con

sideringon

lynecessarycond

ition

may

lead

toinfeasiblesolutio

nOut-of-o

rder

prob

lem

dueto

multip

athrouting

[70]

SCA

PHM

(wo

accumulation)

80211

Internet

No

MP(M

-GW)

L25forw

arding

ge1

Yes

Minim

izem

axlin

kutilizatio

nNo

Radioor

interfe

rencec

onstr

aintsa

reno

taccoun

tedin

initialtraffi

cloadestim

ation

Assum

esam

etrafficd

emands

inallm

eshrouters

Out-of-o

rder

prob

lem

duetomultip

athrouting

Link

rateassig

nmentisc

onsid

ered

[31]

DCA

PRMIRM

Internet

Yes

MP(M

-GW)

Any

ForM

Rswtr

affic

Maxim

izea

ggregated

throug

hput

Min-120582

Out-of-o

rder

prob

lem

andconfi

guratio

ncomplexity

onroutingEa

chradiocanbe

operated

onas

ubseto

fchann

elsC

ombinatio

nof

multip

lechannelsto

singlec

hann

elispo

ssible

[32]

SCA

THM

Internet

Yes

SPAny

Yes

Metric

onho

pcoun

tampinterfe

rence

No

Solvingsubp

roblem

with

localcon

straintsdo

esno

tguarantee

glob

alop


received

power

atan

odew

henanothern

odeis

transm

ittingisassumed

available

[62]

DCA

IRMP

RM

PHM

P2P

No

SPM

PAny

No

Maxim

izea

ggregated

throug

hput

Max-to

-Min


atmessrou

tersIn

single-path

routingop

timality

isno

tguaranteed

since

mathematicalform

ulationisform

ultip

ath

[61]

DCA

B-IRM

P2P

Yes

MP

Any

No

Maxim

ize120582

(scalin

gfactor)

Min-120582

Theq

ualityof

thes

olutionishigh

lydepend

ento

nthep

erform

ance

ofcriticalM

ISfin

ding

algorithm

[33]

SCA

B-IRM

80211

Internet

Lower

ampup

per

boun

dMP(M

GW)

Any

ForM

Rswtr

affic

Maxim

izea

ggregated

throug

hput

Boun

dedfairn

ess

Develo

pedforp

lann

ingsta

gewith

different

resourcesc

onstr

aintsNot

practic

alforlarge

networks

since

fitnessfunctio

ninvolves

solving

entiren

etworkLP

[29]

SCA-

LSTH

MP2

PYes

MP(predefin

ed)

1Yes

Maxim

izea

ggrachieved

demands

No

Link

loadsu

sedto

theC

Aaree

stimated

with

out

consideringradioor

interfe

rencec

onstr

aintsLess

channeld

iversitydu

etolin

kmapping


Table2Con

tinued

Work

Link

layera

ndchannelassignm

ent

Traffi

camprouting

Topo

logy

Objectiv

eampfairn

ess

Noteslimitatio

nsCA

type

Interfe

rence

Techno

logy

Traffi

ctype

Traffi

cload

Routingscheme

Mapping

(1-to-119909)119909

Con

nectivity

Objectiv

eFairn

ess

[37]

SCA

EMInternet

Num

bero

factiv

eflow

sMP(SPper

follo

w)

1Yes

Minim

izec

ontention

(CA)

No

Link

loadse

stimated

with

outcon

sideringradioor

interfe

rencec

onstr

aintsNot

clearlysta

tedwhere

andho

wto

implem

entfl

owratecontrol

[63]

SCA

B-IRM

80211

Internet

Yessequ

ence

oftraffi

cMP

Any

ForM

Rswtr

affic

Minim

izethe

maxim

umcongestio

n(ofw

orst

case)

No

Num

bero

ffacetsincreasee

xpon

entia

llyin

large-scalen


then

umbero

fconstraintsinLP

prob

lemhow

everrelaxingthe

convex

hullto

boxrang

esaffectsthee

fficiency

NoteSCA

static

channelassignm

entDCA

dyn

amicchannelassignm

entLS

link

schedu

lingIM

interfe

rencem

odel

IRMin

terfe

rence-rangem

odel

PRMprotocolm

odel

PHMphysic

almod

elB-bidire

ctional

THMtwo-ho

pmod

elEM

empiric

almeasurementP2

Ppeer-to

-peerSP

single-pathM

Pmultip

athMGWm

ultig

atew

ayM

Rmeshrouterand

LLlinklayer



31 C-JCAR Proposals








Table3State-aw

arec

hann

elassig

nmentrou

tingprop

osalssum

mary

Work

Statea

ware

Approach

Noteslimitatio

nsCh

annelreassignm

ent

Flow

rerouting

[35]

Costisb

ased

onchannel-link

reassig

nment

Not

applicablesin

ceafullyconn

ected

topo

logy

andfix

edpredefinedpaths

arec

onsid

ered

(i)Re

orderc

hann

elin

then

ewCA

toredu

cethec

ost

durin

gmigratio

nfro

mexistingclu

stertothen

ewclu

ster

(i)Out-of-o

rder

prob

lem

(ii)A

ssigning

channelsto

alllogicallin

ksaffectsthe

channeld

iversity

(iii)Perfo

rmance


sequ

encesa

rehigh

lyun

correlated

[85]

Costisb

ased

onchannel-link

reassig

nment

Not

applicablesin

ceafullyconn

ected

topo

logy

isconsidered

(i)Previous

channelsarer

eassignedto

links

ifthey

area

mon

gthes

etof

channelswhich

bestredu

cethe

interfe

rencea

nddo


(i)Internettraffi


atthe

gateway

(ii)P

erform

ance

degrades

with

smalln

umbero

fradios

duetorip

plee

ffecton

channelreassignm

ent

(iii)Assigning

channelsto

alllogicallin

ksaffects

channeld

iversity

[29]

Costisb

ased

onchannel-link

reassig

nment

Not

applicablesin

ceafullyconn

ected

topo

logy

andpredefinedpathsa

reconsidered

(i)Eff

ectiv

edatafor

each

channelisc

ompu

tedand

channelw

iththeh

ighesteffectiv

ecapacity

isselected

Ifthec

hann

elisno

tpreviou

slyassig

nedto

thatlin

kreconfi

guratio

ncostissubtracted

from

thee

ffective

datavalue

(i)Out-of-o

rder

prob

lem

duetomultip

ath

(ii)A

ssigning

channelsto

alllogicallin

ksaffects

channeld

iversity

(iii)Th

echann

elassig

nedto

alinkwill

beconsidered

asfin

alassig

nmentthusearliera

ssignm

entsaren

otaw

areo

flater

assig

nments

thismay

lead

toan

improp

erfin

alchannelassignm

ent

(iv)D

ueto

radioconstraint

somelinks

will

beforced

toassig

nmorec

ongeste


result

inineffi

cientchann

elassig

nment

[84]

Costisb

ased

onchannel-n

ode

reassig

nment

Not

addressed

(i)Startsfro

mthee

xisting

CAandfin

dsnewCA

bymakingsomea

djustm

ento

nthep

reviou

sone

(ii)N

umbero

fradio-chann

elreassig

nmentsis

constrainedto

apredefin

ednu

mber

(i)Perfo

rmance

degrades

with

smalln

umbero

fradios

duetorip

plee

ffecton

channelreassignm

ent

(ii)T


solved

with

CAlin

ksto

optim

izethe

networkperfo

rmance

[37]

Costisb

ased

onchannel-link

reassig

nment

Not

addressed

(i)Startin

gfro


entan

ewchannelisa

ssignedto

alinkifitdo

esno

tviolatethe

radioconstraintsa

ndredu

cesthe

interfe

rence

(ii)N

umbero


entsis

constrainedto

apredefin

ednu

mber

(i)Perfo

rmance

degrades

with

smalln

umbero

fradio

interfa

cesa

nddu

etothatreassig

nmentisrestricted

tochannelsthatdo

notvioletthe

radioconstraints

(ii)A

ssigning

channelsto

alllogicallin

ksaffects

channeld

iversity

(iii)Flow

-levelrou

tingintro

duceso

verheadto

the

routingprocedurea

ndrequ

iresa

routingentryof

each

flowin

routingtablesTh

isisim

practic

alfor

high

numbero

fflow

s

[58]

Not

addressed

Con

sideredw

here

reroutingcostis

basedon

then

umbero

fflow

shaving

different

routingpath

(i)Heuris

ticroutingbasedon

Dijk

straa

lgorith

m(ii)Th

enum

bero

fflow

swhich

allowed

having

new

routes

isconstrainedto

apredefin

edvalue

(i)Flow

reroutingcostisno

taccurately

defin

edFor

insta

ncetwonewflo

wpath

assig

nmentsmay

result

indifferent

disrup

tionwhilebo

tharec

onsid

ered

equally

inreroutingcost

























































Notations










(119906)ℎ119894




1when 119897

2is

active

Competing Interests


References
















































































































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation










Connectivity

Link mapping

Topology info

Channelassignment

Interferencemodel

Technology

Traffic type

Traffic profile

Traffic collection

Routing scheme

Routing table

Execution (Re-)



Algorithms
















Topology

Approaches Traffic

Routing layer

Link layer





1when



119895 Let 119860 isin 119860

119895 Δ(119860) be the set


sum

119890isinΔ(119860)

119909 (119897) le 120593 forall119895 119860 isin 119860119895 (14)






1198821198972

1198971

= 120573 lowast119875V (119903)

119875V (119906)

forall1198971 1198972isin 119871 amp 119897

1= (119903 119904 119888) 1198972 = (119906 V 119888)

(15)



119892 (1198971) + sum

1198972isin1198711198971ampCh(119897

2)=Ch(119897

1)

119873(1198821198972

1198971

) sdot 119892 (1198972) le 120599 (119897

1)


(16)



3 Related Works



Table2C-

JCARprop

osalssum

mary

Work

Link

layera

ndchannelassignm

ent

Traffi

camprouting

Topo

logy

Objectiv

eampfairn

ess

Noteslimitatio

nsCA

type

Interfe

rence

Techno

logy

Traffi

ctype

Traffi

cload

Routingscheme

Mapping

(1-to-119909)119909

Con

nectivity

Objectiv

eFairn

ess

[19]

SCA

B-IRM

80211

P2P

Yes

SPM

P1

Yes

Maxim

izec

ross-sectio

ngood

-put

No

Optim

ality

isno


kcapacityestim

ationin

aheavy

traffi

cloadLess

channeld

iversitydu

etolin

kmapping

[56]

DCA

SC

A-LS

B-IRM

mdashP2

PYes

MP

01

Yes

Maxim

ize120582

(scalin

gfactor)

Min-120582

Out-of-o

rder

prob

lem

andconfi

guratio

ncomplexity

onroutingSw

itching

overhead

issues

[20]

SCA-

LSB-IRM

mdashInternet

Yes

MP

Any

ForM

Rswtr

affic

Maxim

ize120582

(scalin

gfactor)

Min-120582

Radioconstraintsa

reno

tcon

sidered

intheL

POut-of-o

rder

prob

lem

duetomultip

athrouting

[36]

DCA

B-IRM

mdashP2

Psrc-desM

RsMP

Any

No

Maxim

izea

ggregated

throug

hput

No

Out-of-o

rder

prob

lem

andconfi

guratio

ncomplexity

onroutingSw

itching

overhead

issues

Optim

ality

issue

since

onlysufficientcon

ditio

nsarec

onsid

ered

[57]

SCA

B-IRM

80211

P2PInternet

Yes

SPAny

ForM

Rswtr

affic

Maxim

izem

insparelink

capacity

Min-120582

onlin

kutilizatio

n

Usin

gsufficientcon

ditio

nin

MILPmay

notlead

toglob

alop

timalsolutio

nSing

lechannelinitia

ltopo

logy

may

lead

toinfeasiblesolutio

nin

heavy

traffi

cloadscenarios

[30]

SCA-

LSTH

MInternet

Yes

MPpredefined

01

ForM

Rswtr

affic

Maxim

ize120582

(scalin

gfactor)

Min-120582

Con

sideringon

lynecessarycond

ition

may

lead

toinfeasiblesolutio

nOut-of-o

rder

prob

lem

dueto

multip

athrouting

[70]

SCA

PHM

(wo

accumulation)

80211

Internet

No

MP(M

-GW)

L25forw

arding

ge1

Yes

Minim

izem

axlin

kutilizatio

nNo

Radioor

interfe

rencec

onstr

aintsa

reno

taccoun

tedin

initialtraffi

cloadestim

ation

Assum

esam

etrafficd

emands

inallm

eshrouters

Out-of-o

rder

prob

lem

duetomultip

athrouting

Link

rateassig

nmentisc

onsid

ered

[31]

DCA

PRMIRM

Internet

Yes

MP(M

-GW)

Any

ForM

Rswtr

affic

Maxim

izea

ggregated

throug

hput

Min-120582

Out-of-o

rder

prob

lem

andconfi

guratio

ncomplexity

onroutingEa

chradiocanbe

operated

onas

ubseto

fchann

elsC

ombinatio

nof

multip

lechannelsto

singlec

hann

elispo

ssible

[32]

SCA

THM

Internet

Yes

SPAny

Yes

Metric

onho

pcoun

tampinterfe

rence

No

Solvingsubp

roblem

with

localcon

straintsdo

esno

tguarantee

glob

alop


received

power

atan

odew

henanothern

odeis

transm

ittingisassumed

available

[62]

DCA

IRMP

RM

PHM

P2P

No

SPM

PAny

No

Maxim

izea

ggregated

throug

hput

Max-to

-Min


atmessrou

tersIn

single-path

routingop

timality

isno

tguaranteed

since

mathematicalform

ulationisform

ultip

ath

[61]

DCA

B-IRM

P2P

Yes

MP

Any

No

Maxim

ize120582

(scalin

gfactor)

Min-120582

Theq

ualityof

thes

olutionishigh

lydepend

ento

nthep

erform

ance

ofcriticalM

ISfin

ding

algorithm

[33]

SCA

B-IRM

80211

Internet

Lower

ampup

per

boun

dMP(M

GW)

Any

ForM

Rswtr

affic

Maxim

izea

ggregated

throug

hput

Boun

dedfairn

ess

Develo

pedforp

lann

ingsta

gewith

different

resourcesc

onstr

aintsNot

practic

alforlarge

networks

since

fitnessfunctio

ninvolves

solving

entiren

etworkLP

[29]

SCA-

LSTH

MP2

PYes

MP(predefin

ed)

1Yes

Maxim

izea

ggrachieved

demands

No

Link

loadsu

sedto

theC

Aaree

stimated

with

out

consideringradioor

interfe

rencec

onstr

aintsLess

channeld

iversitydu

etolin

kmapping


Table2Con

tinued

Work

Link

layera

ndchannelassignm

ent

Traffi

camprouting

Topo

logy

Objectiv

eampfairn

ess

Noteslimitatio

nsCA

type

Interfe

rence

Techno

logy

Traffi

ctype

Traffi

cload

Routingscheme

Mapping

(1-to-119909)119909

Con

nectivity

Objectiv

eFairn

ess

[37]

SCA

EMInternet

Num

bero

factiv

eflow

sMP(SPper

follo

w)

1Yes

Minim

izec

ontention

(CA)

No

Link

loadse

stimated

with

outcon

sideringradioor

interfe

rencec

onstr

aintsNot

clearlysta

tedwhere

andho

wto

implem

entfl

owratecontrol

[63]

SCA

B-IRM

80211

Internet

Yessequ

ence

oftraffi

cMP

Any

ForM

Rswtr

affic

Minim

izethe

maxim

umcongestio

n(ofw

orst

case)

No

Num

bero

ffacetsincreasee

xpon

entia

llyin

large-scalen


then

umbero

fconstraintsinLP

prob

lemhow

everrelaxingthe

convex

hullto

boxrang

esaffectsthee

fficiency

NoteSCA

static

channelassignm

entDCA

dyn

amicchannelassignm

entLS

link

schedu

lingIM

interfe

rencem

odel

IRMin

terfe

rence-rangem

odel

PRMprotocolm

odel

PHMphysic

almod

elB-bidire

ctional

THMtwo-ho

pmod

elEM

empiric

almeasurementP2

Ppeer-to

-peerSP

single-pathM

Pmultip

athMGWm

ultig

atew

ayM

Rmeshrouterand

LLlinklayer



31 C-JCAR Proposals








Table3State-aw

arec

hann

elassig

nmentrou

tingprop

osalssum

mary

Work

Statea

ware

Approach

Noteslimitatio

nsCh

annelreassignm

ent

Flow

rerouting

[35]

Costisb

ased

onchannel-link

reassig

nment

Not

applicablesin

ceafullyconn

ected

topo

logy

andfix

edpredefinedpaths

arec

onsid

ered

(i)Re

orderc

hann

elin

then

ewCA

toredu

cethec

ost

durin

gmigratio

nfro

mexistingclu

stertothen

ewclu

ster

(i)Out-of-o

rder

prob

lem

(ii)A

ssigning

channelsto

alllogicallin

ksaffectsthe

channeld

iversity

(iii)Perfo

rmance


sequ

encesa

rehigh

lyun

correlated

[85]

Costisb

ased

onchannel-link

reassig

nment

Not

applicablesin

ceafullyconn

ected

topo

logy

isconsidered

(i)Previous

channelsarer

eassignedto

links

ifthey

area

mon

gthes

etof

channelswhich

bestredu

cethe

interfe

rencea

nddo


(i)Internettraffi


atthe

gateway

(ii)P

erform

ance

degrades

with

smalln

umbero

fradios

duetorip

plee

ffecton

channelreassignm

ent

(iii)Assigning

channelsto

alllogicallin

ksaffects

channeld

iversity

[29]

Costisb

ased

onchannel-link

reassig

nment

Not

applicablesin

ceafullyconn

ected

topo

logy

andpredefinedpathsa

reconsidered

(i)Eff

ectiv

edatafor

each

channelisc

ompu

tedand

channelw

iththeh

ighesteffectiv

ecapacity

isselected

Ifthec

hann

elisno

tpreviou

slyassig

nedto

thatlin

kreconfi

guratio

ncostissubtracted

from

thee

ffective

datavalue

(i)Out-of-o

rder

prob

lem

duetomultip

ath

(ii)A

ssigning

channelsto

alllogicallin

ksaffects

channeld

iversity

(iii)Th

echann

elassig

nedto

alinkwill

beconsidered

asfin

alassig

nmentthusearliera

ssignm

entsaren

otaw

areo

flater

assig

nments

thismay

lead

toan

improp

erfin

alchannelassignm

ent

(iv)D

ueto

radioconstraint

somelinks

will

beforced

toassig

nmorec

ongeste


result

inineffi

cientchann

elassig

nment

[84]

Costisb

ased

onchannel-n

ode

reassig

nment

Not

addressed

(i)Startsfro

mthee

xisting

CAandfin

dsnewCA

bymakingsomea

djustm

ento

nthep

reviou

sone

(ii)N

umbero

fradio-chann

elreassig

nmentsis

constrainedto

apredefin

ednu

mber

(i)Perfo

rmance

degrades

with

smalln

umbero

fradios

duetorip

plee

ffecton

channelreassignm

ent

(ii)T


solved

with

CAlin

ksto

optim

izethe

networkperfo

rmance

[37]

Costisb

ased

onchannel-link

reassig

nment

Not

addressed

(i)Startin

gfro


entan

ewchannelisa

ssignedto

alinkifitdo

esno

tviolatethe

radioconstraintsa

ndredu

cesthe

interfe

rence

(ii)N

umbero


entsis

constrainedto

apredefin

ednu

mber

(i)Perfo

rmance

degrades

with

smalln

umbero

fradio

interfa

cesa

nddu

etothatreassig

nmentisrestricted

tochannelsthatdo

notvioletthe

radioconstraints

(ii)A

ssigning

channelsto

alllogicallin

ksaffects

channeld

iversity

(iii)Flow

-levelrou

tingintro

duceso

verheadto

the

routingprocedurea

ndrequ

iresa

routingentryof

each

flowin

routingtablesTh

isisim

practic

alfor

high

numbero

fflow

s

[58]

Not

addressed

Con

sideredw

here

reroutingcostis

basedon

then

umbero

fflow

shaving

different

routingpath

(i)Heuris

ticroutingbasedon

Dijk

straa

lgorith

m(ii)Th

enum

bero

fflow

swhich

allowed

having

new

routes

isconstrainedto

apredefin

edvalue

(i)Flow

reroutingcostisno

taccurately

defin

edFor

insta

ncetwonewflo

wpath

assig

nmentsmay

result

indifferent

disrup

tionwhilebo

tharec

onsid

ered

equally

inreroutingcost

























































Notations










(119906)ℎ119894




1when 119897

2is

active

Competing Interests


References
















































































































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation










Table2C-

JCARprop

osalssum

mary

Work

Link

layera

ndchannelassignm

ent

Traffi

camprouting

Topo

logy

Objectiv

eampfairn

ess

Noteslimitatio

nsCA

type

Interfe

rence

Techno

logy

Traffi

ctype

Traffi

cload

Routingscheme

Mapping

(1-to-119909)119909

Con

nectivity

Objectiv

eFairn

ess

[19]

SCA

B-IRM

80211

P2P

Yes

SPM

P1

Yes

Maxim

izec

ross-sectio

ngood

-put

No

Optim

ality

isno


kcapacityestim

ationin

aheavy

traffi

cloadLess

channeld

iversitydu

etolin

kmapping

[56]

DCA

SC

A-LS

B-IRM

mdashP2

PYes

MP

01

Yes

Maxim

ize120582

(scalin

gfactor)

Min-120582

Out-of-o

rder

prob

lem

andconfi

guratio

ncomplexity

onroutingSw

itching

overhead

issues

[20]

SCA-

LSB-IRM

mdashInternet

Yes

MP

Any

ForM

Rswtr

affic

Maxim

ize120582

(scalin

gfactor)

Min-120582

Radioconstraintsa

reno

tcon

sidered

intheL

POut-of-o

rder

prob

lem

duetomultip

athrouting

[36]

DCA

B-IRM

mdashP2

Psrc-desM

RsMP

Any

No

Maxim

izea

ggregated

throug

hput

No

Out-of-o

rder

prob

lem

andconfi

guratio

ncomplexity

onroutingSw

itching

overhead

issues

Optim

ality

issue

since

onlysufficientcon

ditio

nsarec

onsid

ered

[57]

SCA

B-IRM

80211

P2PInternet

Yes

SPAny

ForM

Rswtr

affic

Maxim

izem

insparelink

capacity

Min-120582

onlin

kutilizatio

n

Usin

gsufficientcon

ditio

nin

MILPmay

notlead

toglob

alop

timalsolutio

nSing

lechannelinitia

ltopo

logy

may

lead

toinfeasiblesolutio

nin

heavy

traffi

cloadscenarios

[30]

SCA-

LSTH

MInternet

Yes

MPpredefined

01

ForM

Rswtr

affic

Maxim

ize120582

(scalin

gfactor)

Min-120582

Con

sideringon

lynecessarycond

ition

may

lead

toinfeasiblesolutio

nOut-of-o

rder

prob

lem

dueto

multip

athrouting

[70]

SCA

PHM

(wo

accumulation)

80211

Internet

No

MP(M

-GW)

L25forw

arding

ge1

Yes

Minim

izem

axlin

kutilizatio

nNo

Radioor

interfe

rencec

onstr

aintsa

reno

taccoun

tedin

initialtraffi

cloadestim

ation

Assum

esam

etrafficd

emands

inallm

eshrouters

Out-of-o

rder

prob

lem

duetomultip

athrouting

Link

rateassig

nmentisc

onsid

ered

[31]

DCA

PRMIRM

Internet

Yes

MP(M

-GW)

Any

ForM

Rswtr

affic

Maxim

izea

ggregated

throug

hput

Min-120582

Out-of-o

rder

prob

lem

andconfi

guratio

ncomplexity

onroutingEa

chradiocanbe

operated

onas

ubseto

fchann

elsC

ombinatio

nof

multip

lechannelsto

singlec

hann

elispo

ssible

[32]

SCA

THM

Internet

Yes

SPAny

Yes

Metric

onho

pcoun

tampinterfe

rence

No

Solvingsubp

roblem

with

localcon

straintsdo

esno

tguarantee

glob

alop


received

power

atan

odew

henanothern

odeis

transm

ittingisassumed

available

[62]

DCA

IRMP

RM

PHM

P2P

No

SPM

PAny

No

Maxim

izea

ggregated

throug

hput

Max-to

-Min


atmessrou

tersIn

single-path

routingop

timality

isno

tguaranteed

since

mathematicalform

ulationisform

ultip

ath

[61]

DCA

B-IRM

P2P

Yes

MP

Any

No

Maxim

ize120582

(scalin

gfactor)

Min-120582

Theq

ualityof

thes

olutionishigh

lydepend

ento

nthep

erform

ance

ofcriticalM

ISfin

ding

algorithm

[33]

SCA

B-IRM

80211

Internet

Lower

ampup

per

boun

dMP(M

GW)

Any

ForM

Rswtr

affic

Maxim

izea

ggregated

throug

hput

Boun

dedfairn

ess

Develo

pedforp

lann

ingsta

gewith

different

resourcesc

onstr

aintsNot

practic

alforlarge

networks

since

fitnessfunctio

ninvolves

solving

entiren

etworkLP

[29]

SCA-

LSTH

MP2

PYes

MP(predefin

ed)

1Yes

Maxim

izea

ggrachieved

demands

No

Link

loadsu

sedto

theC

Aaree

stimated

with

out

consideringradioor

interfe

rencec

onstr

aintsLess

channeld

iversitydu

etolin

kmapping


Table2Con

tinued

Work

Link

layera

ndchannelassignm

ent

Traffi

camprouting

Topo

logy

Objectiv

eampfairn

ess

Noteslimitatio

nsCA

type

Interfe

rence

Techno

logy

Traffi

ctype

Traffi

cload

Routingscheme

Mapping

(1-to-119909)119909

Con

nectivity

Objectiv

eFairn

ess

[37]

SCA

EMInternet

Num

bero

factiv

eflow

sMP(SPper

follo

w)

1Yes

Minim

izec

ontention

(CA)

No

Link

loadse

stimated

with

outcon

sideringradioor

interfe

rencec

onstr

aintsNot

clearlysta

tedwhere

andho

wto

implem

entfl

owratecontrol

[63]

SCA

B-IRM

80211

Internet

Yessequ

ence

oftraffi

cMP

Any

ForM

Rswtr

affic

Minim

izethe

maxim

umcongestio

n(ofw

orst

case)

No

Num

bero

ffacetsincreasee

xpon

entia

llyin

large-scalen


then

umbero

fconstraintsinLP

prob

lemhow

everrelaxingthe

convex

hullto

boxrang

esaffectsthee

fficiency

NoteSCA

static

channelassignm

entDCA

dyn

amicchannelassignm

entLS

link

schedu

lingIM

interfe

rencem

odel

IRMin

terfe

rence-rangem

odel

PRMprotocolm

odel

PHMphysic

almod

elB-bidire

ctional

THMtwo-ho

pmod

elEM

empiric

almeasurementP2

Ppeer-to

-peerSP

single-pathM

Pmultip

athMGWm

ultig

atew

ayM

Rmeshrouterand

LLlinklayer



31 C-JCAR Proposals








Table3State-aw

arec

hann

elassig

nmentrou

tingprop

osalssum

mary

Work

Statea

ware

Approach

Noteslimitatio

nsCh

annelreassignm

ent

Flow

rerouting

[35]

Costisb

ased

onchannel-link

reassig

nment

Not

applicablesin

ceafullyconn

ected

topo

logy

andfix

edpredefinedpaths

arec

onsid

ered

(i)Re

orderc

hann

elin

then

ewCA

toredu

cethec

ost

durin

gmigratio

nfro

mexistingclu

stertothen

ewclu

ster

(i)Out-of-o

rder

prob

lem

(ii)A

ssigning

channelsto

alllogicallin

ksaffectsthe

channeld

iversity

(iii)Perfo

rmance


sequ

encesa

rehigh

lyun

correlated

[85]

Costisb

ased

onchannel-link

reassig

nment

Not

applicablesin

ceafullyconn

ected

topo

logy

isconsidered

(i)Previous

channelsarer

eassignedto

links

ifthey

area

mon

gthes

etof

channelswhich

bestredu

cethe

interfe

rencea

nddo


(i)Internettraffi


atthe

gateway

(ii)P

erform

ance

degrades

with

smalln

umbero

fradios

duetorip

plee

ffecton

channelreassignm

ent

(iii)Assigning

channelsto

alllogicallin

ksaffects

channeld

iversity

[29]

Costisb

ased

onchannel-link

reassig

nment

Not

applicablesin

ceafullyconn

ected

topo

logy

andpredefinedpathsa

reconsidered

(i)Eff

ectiv

edatafor

each

channelisc

ompu

tedand

channelw

iththeh

ighesteffectiv

ecapacity

isselected

Ifthec

hann

elisno

tpreviou

slyassig

nedto

thatlin

kreconfi

guratio

ncostissubtracted

from

thee

ffective

datavalue

(i)Out-of-o

rder

prob

lem

duetomultip

ath

(ii)A

ssigning

channelsto

alllogicallin

ksaffects

channeld

iversity

(iii)Th

echann

elassig

nedto

alinkwill

beconsidered

asfin

alassig

nmentthusearliera

ssignm

entsaren

otaw

areo

flater

assig

nments

thismay

lead

toan

improp

erfin

alchannelassignm

ent

(iv)D

ueto

radioconstraint

somelinks

will

beforced

toassig

nmorec

ongeste


result

inineffi

cientchann

elassig

nment

[84]

Costisb

ased

onchannel-n

ode

reassig

nment

Not

addressed

(i)Startsfro

mthee

xisting

CAandfin

dsnewCA

bymakingsomea

djustm

ento

nthep

reviou

sone

(ii)N

umbero

fradio-chann

elreassig

nmentsis

constrainedto

apredefin

ednu

mber

(i)Perfo

rmance

degrades

with

smalln

umbero

fradios

duetorip

plee

ffecton

channelreassignm

ent

(ii)T


solved

with

CAlin

ksto

optim

izethe

networkperfo

rmance

[37]

Costisb

ased

onchannel-link

reassig

nment

Not

addressed

(i)Startin

gfro


entan

ewchannelisa

ssignedto

alinkifitdo

esno

tviolatethe

radioconstraintsa

ndredu

cesthe

interfe

rence

(ii)N

umbero


entsis

constrainedto

apredefin

ednu

mber

(i)Perfo

rmance

degrades

with

smalln

umbero

fradio

interfa

cesa

nddu

etothatreassig

nmentisrestricted

tochannelsthatdo

notvioletthe

radioconstraints

(ii)A

ssigning

channelsto

alllogicallin

ksaffects

channeld

iversity

(iii)Flow

-levelrou

tingintro

duceso

verheadto

the

routingprocedurea

ndrequ

iresa

routingentryof

each

flowin

routingtablesTh

isisim

practic

alfor

high

numbero

fflow

s

[58]

Not

addressed

Con

sideredw

here

reroutingcostis

basedon

then

umbero

fflow

shaving

different

routingpath

(i)Heuris

ticroutingbasedon

Dijk

straa

lgorith

m(ii)Th

enum

bero

fflow

swhich

allowed

having

new

routes

isconstrainedto

apredefin

edvalue

(i)Flow

reroutingcostisno

taccurately

defin

edFor

insta

ncetwonewflo

wpath

assig

nmentsmay

result

indifferent

disrup

tionwhilebo

tharec

onsid

ered

equally

inreroutingcost

























































Notations










(119906)ℎ119894




1when 119897

2is

active

Competing Interests


References
















































































































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation










Table2Con

tinued

Work

Link

layera

ndchannelassignm

ent

Traffi

camprouting

Topo

logy

Objectiv

eampfairn

ess

Noteslimitatio

nsCA

type

Interfe

rence

Techno

logy

Traffi

ctype

Traffi

cload

Routingscheme

Mapping

(1-to-119909)119909

Con

nectivity

Objectiv

eFairn

ess

[37]

SCA

EMInternet

Num

bero

factiv

eflow

sMP(SPper

follo

w)

1Yes

Minim

izec

ontention

(CA)

No

Link

loadse

stimated

with

outcon

sideringradioor

interfe

rencec

onstr

aintsNot

clearlysta

tedwhere

andho

wto

implem

entfl

owratecontrol

[63]

SCA

B-IRM

80211

Internet

Yessequ

ence

oftraffi

cMP

Any

ForM

Rswtr

affic

Minim

izethe

maxim

umcongestio

n(ofw

orst

case)

No

Num

bero

ffacetsincreasee

xpon

entia

llyin

large-scalen


then

umbero

fconstraintsinLP

prob

lemhow

everrelaxingthe

convex

hullto

boxrang

esaffectsthee

fficiency

NoteSCA

static

channelassignm

entDCA

dyn

amicchannelassignm

entLS

link

schedu

lingIM

interfe

rencem

odel

IRMin

terfe

rence-rangem

odel

PRMprotocolm

odel

PHMphysic

almod

elB-bidire

ctional

THMtwo-ho

pmod

elEM

empiric

almeasurementP2

Ppeer-to

-peerSP

single-pathM

Pmultip

athMGWm

ultig

atew

ayM

Rmeshrouterand

LLlinklayer



31 C-JCAR Proposals








Table3State-aw

arec

hann

elassig

nmentrou

tingprop

osalssum

mary

Work

Statea

ware

Approach

Noteslimitatio

nsCh

annelreassignm

ent

Flow

rerouting

[35]

Costisb

ased

onchannel-link

reassig

nment

Not

applicablesin

ceafullyconn

ected

topo

logy

andfix

edpredefinedpaths

arec

onsid

ered

(i)Re

orderc

hann

elin

then

ewCA

toredu

cethec

ost

durin

gmigratio

nfro

mexistingclu

stertothen

ewclu

ster

(i)Out-of-o

rder

prob

lem

(ii)A

ssigning

channelsto

alllogicallin

ksaffectsthe

channeld

iversity

(iii)Perfo

rmance


sequ

encesa

rehigh

lyun

correlated

[85]

Costisb

ased

onchannel-link

reassig

nment

Not

applicablesin

ceafullyconn

ected

topo

logy

isconsidered

(i)Previous

channelsarer

eassignedto

links

ifthey

area

mon

gthes

etof

channelswhich

bestredu

cethe

interfe

rencea

nddo


(i)Internettraffi


atthe

gateway

(ii)P

erform

ance

degrades

with

smalln

umbero

fradios

duetorip

plee

ffecton

channelreassignm

ent

(iii)Assigning

channelsto

alllogicallin

ksaffects

channeld

iversity

[29]

Costisb

ased

onchannel-link

reassig

nment

Not

applicablesin

ceafullyconn

ected

topo

logy

andpredefinedpathsa

reconsidered

(i)Eff

ectiv

edatafor

each

channelisc

ompu

tedand

channelw

iththeh

ighesteffectiv

ecapacity

isselected

Ifthec

hann

elisno

tpreviou

slyassig

nedto

thatlin

kreconfi

guratio

ncostissubtracted

from

thee

ffective

datavalue

(i)Out-of-o

rder

prob

lem

duetomultip

ath

(ii)A

ssigning

channelsto

alllogicallin

ksaffects

channeld

iversity

(iii)Th

echann

elassig

nedto

alinkwill

beconsidered

asfin

alassig

nmentthusearliera

ssignm

entsaren

otaw

areo

flater

assig

nments

thismay

lead

toan

improp

erfin

alchannelassignm

ent

(iv)D

ueto

radioconstraint

somelinks

will

beforced

toassig

nmorec

ongeste


result

inineffi

cientchann

elassig

nment

[84]

Costisb

ased

onchannel-n

ode

reassig

nment

Not

addressed

(i)Startsfro

mthee

xisting

CAandfin

dsnewCA

bymakingsomea

djustm

ento

nthep

reviou

sone

(ii)N

umbero

fradio-chann

elreassig

nmentsis

constrainedto

apredefin

ednu

mber

(i)Perfo

rmance

degrades

with

smalln

umbero

fradios

duetorip

plee

ffecton

channelreassignm

ent

(ii)T


solved

with

CAlin

ksto

optim

izethe

networkperfo

rmance

[37]

Costisb

ased

onchannel-link

reassig

nment

Not

addressed

(i)Startin

gfro


entan

ewchannelisa

ssignedto

alinkifitdo

esno

tviolatethe

radioconstraintsa

ndredu

cesthe

interfe

rence

(ii)N

umbero


entsis

constrainedto

apredefin

ednu

mber

(i)Perfo

rmance

degrades

with

smalln

umbero

fradio

interfa

cesa

nddu

etothatreassig

nmentisrestricted

tochannelsthatdo

notvioletthe

radioconstraints

(ii)A

ssigning

channelsto

alllogicallin

ksaffects

channeld

iversity

(iii)Flow

-levelrou

tingintro

duceso

verheadto

the

routingprocedurea

ndrequ

iresa

routingentryof

each

flowin

routingtablesTh

isisim

practic

alfor

high

numbero

fflow

s

[58]

Not

addressed

Con

sideredw

here

reroutingcostis

basedon

then

umbero

fflow

shaving

different

routingpath

(i)Heuris

ticroutingbasedon

Dijk

straa

lgorith

m(ii)Th

enum

bero

fflow

swhich

allowed

having

new

routes

isconstrainedto

apredefin

edvalue

(i)Flow

reroutingcostisno

taccurately

defin

edFor

insta

ncetwonewflo

wpath

assig

nmentsmay

result

indifferent

disrup

tionwhilebo

tharec

onsid

ered

equally

inreroutingcost

























































Notations










(119906)ℎ119894




1when 119897

2is

active

Competing Interests


References
















































































































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation











31 C-JCAR Proposals








Table3State-aw

arec

hann

elassig

nmentrou

tingprop

osalssum

mary

Work

Statea

ware

Approach

Noteslimitatio

nsCh

annelreassignm

ent

Flow

rerouting

[35]

Costisb

ased

onchannel-link

reassig

nment

Not

applicablesin

ceafullyconn

ected

topo

logy

andfix

edpredefinedpaths

arec

onsid

ered

(i)Re

orderc

hann

elin

then

ewCA

toredu

cethec

ost

durin

gmigratio

nfro

mexistingclu

stertothen

ewclu

ster

(i)Out-of-o

rder

prob

lem

(ii)A

ssigning

channelsto

alllogicallin

ksaffectsthe

channeld

iversity

(iii)Perfo

rmance


sequ

encesa

rehigh

lyun

correlated

[85]

Costisb

ased

onchannel-link

reassig

nment

Not

applicablesin

ceafullyconn

ected

topo

logy

isconsidered

(i)Previous

channelsarer

eassignedto

links

ifthey

area

mon

gthes

etof

channelswhich

bestredu

cethe

interfe

rencea

nddo


(i)Internettraffi


atthe

gateway

(ii)P

erform

ance

degrades

with

smalln

umbero

fradios

duetorip

plee

ffecton

channelreassignm

ent

(iii)Assigning

channelsto

alllogicallin

ksaffects

channeld

iversity

[29]

Costisb

ased

onchannel-link

reassig

nment

Not

applicablesin

ceafullyconn

ected

topo

logy

andpredefinedpathsa

reconsidered

(i)Eff

ectiv

edatafor

each

channelisc

ompu

tedand

channelw

iththeh

ighesteffectiv

ecapacity

isselected

Ifthec

hann

elisno

tpreviou

slyassig

nedto

thatlin

kreconfi

guratio

ncostissubtracted

from

thee

ffective

datavalue

(i)Out-of-o

rder

prob

lem

duetomultip

ath

(ii)A

ssigning

channelsto

alllogicallin

ksaffects

channeld

iversity

(iii)Th

echann

elassig

nedto

alinkwill

beconsidered

asfin

alassig

nmentthusearliera

ssignm

entsaren

otaw

areo

flater

assig

nments

thismay

lead

toan

improp

erfin

alchannelassignm

ent

(iv)D

ueto

radioconstraint

somelinks

will

beforced

toassig

nmorec

ongeste


result

inineffi

cientchann

elassig

nment

[84]

Costisb

ased

onchannel-n

ode

reassig

nment

Not

addressed

(i)Startsfro

mthee

xisting

CAandfin

dsnewCA

bymakingsomea

djustm

ento

nthep

reviou

sone

(ii)N

umbero

fradio-chann

elreassig

nmentsis

constrainedto

apredefin

ednu

mber

(i)Perfo

rmance

degrades

with

smalln

umbero

fradios

duetorip

plee

ffecton

channelreassignm

ent

(ii)T


solved

with

CAlin

ksto

optim

izethe

networkperfo

rmance

[37]

Costisb

ased

onchannel-link

reassig

nment

Not

addressed

(i)Startin

gfro


entan

ewchannelisa

ssignedto

alinkifitdo

esno

tviolatethe

radioconstraintsa

ndredu

cesthe

interfe

rence

(ii)N

umbero


entsis

constrainedto

apredefin

ednu

mber

(i)Perfo

rmance

degrades

with

smalln

umbero

fradio

interfa

cesa

nddu

etothatreassig

nmentisrestricted

tochannelsthatdo

notvioletthe

radioconstraints

(ii)A

ssigning

channelsto

alllogicallin

ksaffects

channeld

iversity

(iii)Flow

-levelrou

tingintro

duceso

verheadto

the

routingprocedurea

ndrequ

iresa

routingentryof

each

flowin

routingtablesTh

isisim

practic

alfor

high

numbero

fflow

s

[58]

Not

addressed

Con

sideredw

here

reroutingcostis

basedon

then

umbero

fflow

shaving

different

routingpath

(i)Heuris

ticroutingbasedon

Dijk

straa

lgorith

m(ii)Th

enum

bero

fflow

swhich

allowed

having

new

routes

isconstrainedto

apredefin

edvalue

(i)Flow

reroutingcostisno

taccurately

defin

edFor

insta

ncetwonewflo

wpath

assig

nmentsmay

result

indifferent

disrup

tionwhilebo

tharec

onsid

ered

equally

inreroutingcost

























































Notations










(119906)ℎ119894




1when 119897

2is

active

Competing Interests


References
















































































































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation










Table3State-aw

arec

hann

elassig

nmentrou

tingprop

osalssum

mary

Work

Statea

ware

Approach

Noteslimitatio

nsCh

annelreassignm

ent

Flow

rerouting

[35]

Costisb

ased

onchannel-link

reassig

nment

Not

applicablesin

ceafullyconn

ected

topo

logy

andfix

edpredefinedpaths

arec

onsid

ered

(i)Re

orderc

hann

elin

then

ewCA

toredu

cethec

ost

durin

gmigratio

nfro

mexistingclu

stertothen

ewclu

ster

(i)Out-of-o

rder

prob

lem

(ii)A

ssigning

channelsto

alllogicallin

ksaffectsthe

channeld

iversity

(iii)Perfo

rmance


sequ

encesa

rehigh

lyun

correlated

[85]

Costisb

ased

onchannel-link

reassig

nment

Not

applicablesin

ceafullyconn

ected

topo

logy

isconsidered

(i)Previous

channelsarer

eassignedto

links

ifthey

area

mon

gthes

etof

channelswhich

bestredu

cethe

interfe

rencea

nddo


(i)Internettraffi


atthe

gateway

(ii)P

erform

ance

degrades

with

smalln

umbero

fradios

duetorip

plee

ffecton

channelreassignm

ent

(iii)Assigning

channelsto

alllogicallin

ksaffects

channeld

iversity

[29]

Costisb

ased

onchannel-link

reassig

nment

Not

applicablesin

ceafullyconn

ected

topo

logy

andpredefinedpathsa

reconsidered

(i)Eff

ectiv

edatafor

each

channelisc

ompu

tedand

channelw

iththeh

ighesteffectiv

ecapacity

isselected

Ifthec

hann

elisno

tpreviou

slyassig

nedto

thatlin

kreconfi

guratio

ncostissubtracted

from

thee

ffective

datavalue

(i)Out-of-o

rder

prob

lem

duetomultip

ath

(ii)A

ssigning

channelsto

alllogicallin

ksaffects

channeld

iversity

(iii)Th

echann

elassig

nedto

alinkwill

beconsidered

asfin

alassig

nmentthusearliera

ssignm

entsaren

otaw

areo

flater

assig

nments

thismay

lead

toan

improp

erfin

alchannelassignm

ent

(iv)D

ueto

radioconstraint

somelinks

will

beforced

toassig

nmorec

ongeste


result

inineffi

cientchann

elassig

nment

[84]

Costisb

ased

onchannel-n

ode

reassig

nment

Not

addressed

(i)Startsfro

mthee

xisting

CAandfin

dsnewCA

bymakingsomea

djustm

ento

nthep

reviou

sone

(ii)N

umbero

fradio-chann

elreassig

nmentsis

constrainedto

apredefin

ednu

mber

(i)Perfo

rmance

degrades

with

smalln

umbero

fradios

duetorip

plee

ffecton

channelreassignm

ent

(ii)T


solved

with

CAlin

ksto

optim

izethe

networkperfo

rmance

[37]

Costisb

ased

onchannel-link

reassig

nment

Not

addressed

(i)Startin

gfro


entan

ewchannelisa

ssignedto

alinkifitdo

esno

tviolatethe

radioconstraintsa

ndredu

cesthe

interfe

rence

(ii)N

umbero


entsis

constrainedto

apredefin

ednu

mber

(i)Perfo

rmance

degrades

with

smalln

umbero

fradio

interfa

cesa

nddu

etothatreassig

nmentisrestricted

tochannelsthatdo

notvioletthe

radioconstraints

(ii)A

ssigning

channelsto

alllogicallin

ksaffects

channeld

iversity

(iii)Flow

-levelrou

tingintro

duceso

verheadto

the

routingprocedurea

ndrequ

iresa

routingentryof

each

flowin

routingtablesTh

isisim

practic

alfor

high

numbero

fflow

s

[58]

Not

addressed

Con

sideredw

here

reroutingcostis

basedon

then

umbero

fflow

shaving

different

routingpath

(i)Heuris

ticroutingbasedon

Dijk

straa

lgorith

m(ii)Th

enum

bero

fflow

swhich

allowed

having

new

routes

isconstrainedto

apredefin

edvalue

(i)Flow

reroutingcostisno

taccurately

defin

edFor

insta

ncetwonewflo

wpath

assig

nmentsmay

result

indifferent

disrup

tionwhilebo

tharec

onsid

ered

equally

inreroutingcost

























































Notations










(119906)ℎ119894




1when 119897

2is

active

Competing Interests


References
















































































































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation

































































Notations










(119906)ℎ119894




1when 119897

2is

active

Competing Interests


References
















































































































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation



















































































































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation









review article joint channel assignment and routing in multiradio...

Documents