Distributed Channel Assignment and Routing Multiradio Mutlichannel Multihop Wireless Networks Haitao...

33
Distributed Channel Assignment Distributed Channel Assignment and Routing Multiradio Mutlicha and Routing Multiradio Mutlicha nnel Multihop Wireless Networks nnel Multihop Wireless Networks Haitao Wu, Fan Yang, Kun Tan, Jie Chen, Qian Haitao Wu, Fan Yang, Kun Tan, Jie Chen, Qian Zhang, and Zhenshrng Zhang Zhang, and Zhenshrng Zhang Journal on Selected Areas in Communications 2006 Journal on Selected Areas in Communications 2006

Transcript of Distributed Channel Assignment and Routing Multiradio Mutlichannel Multihop Wireless Networks Haitao...

Page 1: Distributed Channel Assignment and Routing Multiradio Mutlichannel Multihop Wireless Networks Haitao Wu, Fan Yang, Kun Tan, Jie Chen, Qian Zhang, and Zhenshrng.

Distributed Channel Assignment and RoutiDistributed Channel Assignment and Routing Multiradio Mutlichannel Multihop Wireleng Multiradio Mutlichannel Multihop Wirele

ss Networksss Networks

Haitao Wu, Fan Yang, Kun Tan, Jie Chen, Qian Zhang, and Haitao Wu, Fan Yang, Kun Tan, Jie Chen, Qian Zhang, and Zhenshrng ZhangZhenshrng Zhang

Journal on Selected Areas in Communications 2006Journal on Selected Areas in Communications 2006

Page 2: Distributed Channel Assignment and Routing Multiradio Mutlichannel Multihop Wireless Networks Haitao Wu, Fan Yang, Kun Tan, Jie Chen, Qian Zhang, and Zhenshrng.

OutlineOutline

IntroductionIntroduction Problem StatementProblem Statement CCM: Metric for JCARCCM: Metric for JCAR Distributed Algorithm for JCARDistributed Algorithm for JCAR SimulationsSimulations ConclusionsConclusions

Page 3: Distributed Channel Assignment and Routing Multiradio Mutlichannel Multihop Wireless Networks Haitao Wu, Fan Yang, Kun Tan, Jie Chen, Qian Zhang, and Zhenshrng.

IntroductionIntroduction Multiple orthogonal channels are defined in IEMultiple orthogonal channels are defined in IE

EE standards.(3 for 802.11b/13 for 802.11a)EE standards.(3 for 802.11b/13 for 802.11a)

With cheaper hardware and diverse wireless teWith cheaper hardware and diverse wireless technologies, many mobile devices are equippechnologies, many mobile devices are equipped with multiple radios.d with multiple radios.

These devices may construct a multiradio mulThese devices may construct a multiradio multichannel multihop wireless network.(Mtichannel multihop wireless network.(M33WN)WN)

Page 4: Distributed Channel Assignment and Routing Multiradio Mutlichannel Multihop Wireless Networks Haitao Wu, Fan Yang, Kun Tan, Jie Chen, Qian Zhang, and Zhenshrng.

IntroductionIntroduction

How to effectively leverage multiradio and muHow to effectively leverage multiradio and multichannel to conquer/reduce the wireless inteltichannel to conquer/reduce the wireless interference that widely exists in the classical multrference that widely exists in the classical multihop wireless networks.ihop wireless networks.

To effectively mitigate interference, both routiTo effectively mitigate interference, both routing and channel assignment (CA) should be carng and channel assignment (CA) should be carefully designed.efully designed.

Page 5: Distributed Channel Assignment and Routing Multiradio Mutlichannel Multihop Wireless Networks Haitao Wu, Fan Yang, Kun Tan, Jie Chen, Qian Zhang, and Zhenshrng.

Introduction - ChallengesIntroduction - Challenges

A quantitative measure of the performance A quantitative measure of the performance gain of any new JCAR pattern to denote any gain of any new JCAR pattern to denote any specific combined solution of CA and specific combined solution of CA and routing.routing.

There might be no common radio or There might be no common radio or common channel supported in the whole common channel supported in the whole network for both data transmission and network for both data transmission and signaling.signaling.

The signaling overhead to obtain updated The signaling overhead to obtain updated traffic and link-state information for JCAR in traffic and link-state information for JCAR in such a heterogeneous environment.such a heterogeneous environment.

Page 6: Distributed Channel Assignment and Routing Multiradio Mutlichannel Multihop Wireless Networks Haitao Wu, Fan Yang, Kun Tan, Jie Chen, Qian Zhang, and Zhenshrng.

Introduction - ChallengesIntroduction - Challenges

The overhead of 802.11 a/b/g NICs on chThe overhead of 802.11 a/b/g NICs on channel switching should be taken into accannel switching should be taken into account.ount.

Page 7: Distributed Channel Assignment and Routing Multiradio Mutlichannel Multihop Wireless Networks Haitao Wu, Fan Yang, Kun Tan, Jie Chen, Qian Zhang, and Zhenshrng.

Solution SummariesSolution Summaries

Channel Cost Metric (CCM) reflects Channel Cost Metric (CCM) reflects the expected transmission cost the expected transmission cost weighted by channel utilization.weighted by channel utilization.

““equivalent fraction of air timeequivalent fraction of air time” (FAT) ” (FAT) reflects the channel busy time and reflects the channel busy time and provides a common reference value provides a common reference value for heterogeneous radios.for heterogeneous radios.

Page 8: Distributed Channel Assignment and Routing Multiradio Mutlichannel Multihop Wireless Networks Haitao Wu, Fan Yang, Kun Tan, Jie Chen, Qian Zhang, and Zhenshrng.

Solution SummariesSolution Summaries

A distributed algorithm that effectively A distributed algorithm that effectively selects the JCAR pattern which has the selects the JCAR pattern which has the smallest CCM value among a subset of smallest CCM value among a subset of potential JCAR patterns.potential JCAR patterns.

Page 9: Distributed Channel Assignment and Routing Multiradio Mutlichannel Multihop Wireless Networks Haitao Wu, Fan Yang, Kun Tan, Jie Chen, Qian Zhang, and Zhenshrng.

Problem StatementProblem StatementInitial Channel

Equips two NICs

Flow1 on 11g NIC and channel g1

New Flow

Page 10: Distributed Channel Assignment and Routing Multiradio Mutlichannel Multihop Wireless Networks Haitao Wu, Fan Yang, Kun Tan, Jie Chen, Qian Zhang, and Zhenshrng.

CCM: Metric for JCARCCM: Metric for JCAR

CCM represents the expected per-unit CCM represents the expected per-unit transmission time on each channel weighted by transmission time on each channel weighted by channel utilization and explicitly captures the channel utilization and explicitly captures the effect of the interference from hidden links.effect of the interference from hidden links.

Page 11: Distributed Channel Assignment and Routing Multiradio Mutlichannel Multihop Wireless Networks Haitao Wu, Fan Yang, Kun Tan, Jie Chen, Qian Zhang, and Zhenshrng.

Intuition for Metric DefinitionIntuition for Metric Definition

Expected Transmission Time, Expected Transmission Time, ETTETTlli i is the sum of transis the sum of trans

mission time of all data packets sent on channel mission time of all data packets sent on channel ll over over a time unit.a time unit.

Page 12: Distributed Channel Assignment and Routing Multiradio Mutlichannel Multihop Wireless Networks Haitao Wu, Fan Yang, Kun Tan, Jie Chen, Qian Zhang, and Zhenshrng.

Intuition for Metric DefinitionIntuition for Metric Definition

FAT represents the normalized overall channel FAT represents the normalized overall channel utilization.utilization.

For channel For channel ll at node at node ii, FAT, F, FAT, Fllii, is defined as the , is defined as the

ratio of the total air time consumed.ratio of the total air time consumed.

Page 13: Distributed Channel Assignment and Routing Multiradio Mutlichannel Multihop Wireless Networks Haitao Wu, Fan Yang, Kun Tan, Jie Chen, Qian Zhang, and Zhenshrng.

Intuition for Metric DefinitionIntuition for Metric Definition

Based on this metric, it is apparent that pattern2 is better than pattern1.

Page 14: Distributed Channel Assignment and Routing Multiradio Mutlichannel Multihop Wireless Networks Haitao Wu, Fan Yang, Kun Tan, Jie Chen, Qian Zhang, and Zhenshrng.

Expected Transmission TimeExpected Transmission TimeThe average payload

sizeThe average transmission

time of one data frame

Collision probability

Page 15: Distributed Channel Assignment and Routing Multiradio Mutlichannel Multihop Wireless Networks Haitao Wu, Fan Yang, Kun Tan, Jie Chen, Qian Zhang, and Zhenshrng.

Normalized Channel Occupation Normalized Channel Occupation Time-Fraction of Air TimeTime-Fraction of Air Time

Page 16: Distributed Channel Assignment and Routing Multiradio Mutlichannel Multihop Wireless Networks Haitao Wu, Fan Yang, Kun Tan, Jie Chen, Qian Zhang, and Zhenshrng.

Normalized Channel Occupation TiNormalized Channel Occupation Time-Fraction of Air Timeme-Fraction of Air Time

Page 17: Distributed Channel Assignment and Routing Multiradio Mutlichannel Multihop Wireless Networks Haitao Wu, Fan Yang, Kun Tan, Jie Chen, Qian Zhang, and Zhenshrng.

A Close-Form Expression for CCMA Close-Form Expression for CCM

Page 18: Distributed Channel Assignment and Routing Multiradio Mutlichannel Multihop Wireless Networks Haitao Wu, Fan Yang, Kun Tan, Jie Chen, Qian Zhang, and Zhenshrng.

Distributed Algorithm for JCARDistributed Algorithm for JCAR

A better pattern is found, the system may A better pattern is found, the system may be reconfigured, and the total channel be reconfigured, and the total channel cost is reduced.cost is reduced.

JCAR guarantees this by performing a JCAR guarantees this by performing a feasibility check before applying a new feasibility check before applying a new pattern.pattern.

JCAR reconfigurations are performed JCAR reconfigurations are performed periodically.periodically.

Page 19: Distributed Channel Assignment and Routing Multiradio Mutlichannel Multihop Wireless Networks Haitao Wu, Fan Yang, Kun Tan, Jie Chen, Qian Zhang, and Zhenshrng.

Distributed Algorithm for JCARDistributed Algorithm for JCAR

1.1. Identify some of the possible JCAR patterns and for Identify some of the possible JCAR patterns and for each of the identified JCAR patterns:each of the identified JCAR patterns:

1)1) Check feasibility Check feasibility F F llii (new pattern) <= Th(new pattern) <= ThF F ;;

2)2) Check network connectivity;Check network connectivity;3)3) Determine its CCM value if feasibleDetermine its CCM value if feasible

2.2. Select a feasible pattern which has the smallest CCM Select a feasible pattern which has the smallest CCM value.value.

3.3. Conduct the switching operationConduct the switching operation

Page 20: Distributed Channel Assignment and Routing Multiradio Mutlichannel Multihop Wireless Networks Haitao Wu, Fan Yang, Kun Tan, Jie Chen, Qian Zhang, and Zhenshrng.

JCAR Candidate Pattern SelectionJCAR Candidate Pattern Selection

Only changing channels within the Only changing channels within the channel set to which cannel channel set to which cannel ll belongs belongs

Only changing interface between Only changing interface between nodes nodes ii and and j j

Combination of both channel change Combination of both channel change and interface switchingand interface switching

Page 21: Distributed Channel Assignment and Routing Multiradio Mutlichannel Multihop Wireless Networks Haitao Wu, Fan Yang, Kun Tan, Jie Chen, Qian Zhang, and Zhenshrng.

Feasibility for CA and Routing Feasibility for CA and Routing

A new JCAR pattern is feasible. A new JCAR pattern is feasible.

(no flow suffers degradation in throughput)(no flow suffers degradation in throughput)

Page 22: Distributed Channel Assignment and Routing Multiradio Mutlichannel Multihop Wireless Networks Haitao Wu, Fan Yang, Kun Tan, Jie Chen, Qian Zhang, and Zhenshrng.

Network ConnectivityNetwork Connectivity

Some JCAR patterns may result in network Some JCAR patterns may result in network partitions and these patterns should be partitions and these patterns should be avoid.avoid.

Connectivity Invariance RuleConnectivity Invariance Rule• If any node-pair that was originally connected, the If any node-pair that was originally connected, the

node-pair should still be connected under a new node-pair should still be connected under a new JCAR pattern.JCAR pattern.

Only patterns involved with switching Only patterns involved with switching channels (categories 1 or 3 above) between channels (categories 1 or 3 above) between two nodes may result in network partition.two nodes may result in network partition.

Page 23: Distributed Channel Assignment and Routing Multiradio Mutlichannel Multihop Wireless Networks Haitao Wu, Fan Yang, Kun Tan, Jie Chen, Qian Zhang, and Zhenshrng.

Network ConnectivityNetwork Connectivity

l → l’l

l

Chain puzzle may involve a large number of nodes It is difficult to synchronize the switching actions among all nodes

Page 24: Distributed Channel Assignment and Routing Multiradio Mutlichannel Multihop Wireless Networks Haitao Wu, Fan Yang, Kun Tan, Jie Chen, Qian Zhang, and Zhenshrng.

Network ConnectivityNetwork Connectivity

In the above case, node In the above case, node ii, , jj should give up should give up the channel switching.the channel switching.

The node initiates a switching request will The node initiates a switching request will check if there contains check if there contains chain puzzlechain puzzle within its within its two hop information.two hop information.

Page 25: Distributed Channel Assignment and Routing Multiradio Mutlichannel Multihop Wireless Networks Haitao Wu, Fan Yang, Kun Tan, Jie Chen, Qian Zhang, and Zhenshrng.

Distributed JCARDistributed JCAR

If the CCM value under the new pattern is smaller If the CCM value under the new pattern is smaller than the current one by a predefined threshold THthan the current one by a predefined threshold THCACA, , then node then node ii starts the operation for CA and routing starts the operation for CA and routing switching.switching.

For interface switching, the action is taken locally, For interface switching, the action is taken locally, only between node only between node ii and and jj..

For channel switching, a distributed procedure is For channel switching, a distributed procedure is required to guarantee that all the neighbors accept required to guarantee that all the neighbors accept such changes,.such changes,.

Page 26: Distributed Channel Assignment and Routing Multiradio Mutlichannel Multihop Wireless Networks Haitao Wu, Fan Yang, Kun Tan, Jie Chen, Qian Zhang, and Zhenshrng.

Distributed JCARDistributed JCARthe CCM value is the CCM value is smaller than THsmaller than THCACA

To indicate that it wishes to switch

channel l to l’

To avoid synchronized pattern selection triggered for nodes in a neighborhood

Page 27: Distributed Channel Assignment and Routing Multiradio Mutlichannel Multihop Wireless Networks Haitao Wu, Fan Yang, Kun Tan, Jie Chen, Qian Zhang, and Zhenshrng.

SimulationsSimulations

Using NS-2 simulatorUsing NS-2 simulator A mesh network on a grid with 36 nodes, A mesh network on a grid with 36 nodes, The channel model is TwoRayGroundThe channel model is TwoRayGround Trigger threshold THTrigger threshold THHH is is 0.80.8 Setting to 802.11a DCFSetting to 802.11a DCF with 54Mb/swith 54Mb/s

Page 28: Distributed Channel Assignment and Routing Multiradio Mutlichannel Multihop Wireless Networks Haitao Wu, Fan Yang, Kun Tan, Jie Chen, Qian Zhang, and Zhenshrng.

SimulationsSimulations

Page 29: Distributed Channel Assignment and Routing Multiradio Mutlichannel Multihop Wireless Networks Haitao Wu, Fan Yang, Kun Tan, Jie Chen, Qian Zhang, and Zhenshrng.

SimulationsSimulations

Page 30: Distributed Channel Assignment and Routing Multiradio Mutlichannel Multihop Wireless Networks Haitao Wu, Fan Yang, Kun Tan, Jie Chen, Qian Zhang, and Zhenshrng.

SimulationsSimulations

Another evaluation is in a nine-node wireless testbed.Another evaluation is in a nine-node wireless testbed. All nine nodes are DELLAll nine nodes are DELL

OPTIPLEXOPTIPLEX Nodes 10, 24, 27, 28 Nodes 10, 24, 27, 28

and 31 are equipped and 31 are equipped with two 802.11a/b/gwith two 802.11a/b/g

combo cards.combo cards.

Page 31: Distributed Channel Assignment and Routing Multiradio Mutlichannel Multihop Wireless Networks Haitao Wu, Fan Yang, Kun Tan, Jie Chen, Qian Zhang, and Zhenshrng.

SimulationsSimulations

Page 32: Distributed Channel Assignment and Routing Multiradio Mutlichannel Multihop Wireless Networks Haitao Wu, Fan Yang, Kun Tan, Jie Chen, Qian Zhang, and Zhenshrng.

SimulationsSimulations

Page 33: Distributed Channel Assignment and Routing Multiradio Mutlichannel Multihop Wireless Networks Haitao Wu, Fan Yang, Kun Tan, Jie Chen, Qian Zhang, and Zhenshrng.

ConclusionsConclusions

Discussion how to improve the performance Discussion how to improve the performance by joint CA and routing for a heterogeneous by joint CA and routing for a heterogeneous MM33WN.WN.

Presenting CCM a critical metric that Presenting CCM a critical metric that quantifies the difference for various JCAR quantifies the difference for various JCAR patterns.patterns.

A distributed algorithm, JCAR, is to select a A distributed algorithm, JCAR, is to select a JCAR pattern that has smallest CCM at each JCAR pattern that has smallest CCM at each node locally.node locally.