B. Bellalta Mobile Communication Networks

Transcription

1 IEEE e : EDCA B. Bellalta Mobile Communication Networks

2 Scenario STA AP STA Server Server Fixed Network STA Server Upwnlink TCP flows Downlink TCP flows STA AP STA What is the WLAN cell performance with bidirectional TCP traffic? STA

3 3 VoIP over WLANs Voice over IP Economical alternative to traditional telephony. Wireless LANs August 2003 Broadband access network to internet. VoIP over WLANs. A market opportunity. Complement/Substitute of current cellular networks. Technical limitations. 178,000 new Skype users per day!

4 4 VoIP over WLANs Voice over IP Economical alternative to traditional telephony. Wireless LANs Broadband access network to internet. VoIP over WLANs. A market opportunity. Complement/Substitute of current cellular networks. Technical limitations. August ,000 new Skype users per day!

5 VoIP over WLANs Limitations for VoIPoWLAN (and other real time services) Protocol overheads. Capacity varying channels (multi rate). Random Access MAC. Bandwidth penalty due to the distributed access. Unfairnes Downlink/Uplink. Difficult coexistence between heterogeneous (rigid and elastic) flows. Hand off / Roaming between WLANs. 5

6 6 VoIP over WLANs Example: Rdata = 2 Mbps, Rbasic = 1 Mbps, VoIP codec: G.729 (8 Kbps) Downlink/Uplink TCP flows simultaneous VoIP calls TCP packets packets packets TCP packets VoIP call packets TCP packets VoIP call IEEE Solution: EDCA (Enhanced Distributed Channel Access): Traffic differentiation at MAC layer + CAC optional. EDCA is too conservative. Static Traffic Differentiation without Admission Control.

7 7 Multimedia Traffic Flows General Classification (queue occupation). Elastic flows (TCP based): use all available bandwidth, fair with other elastic flows. saturated source/queue with : B L/X, ρ =1 Streaming/Rigid flows (UDP based): use only the bandwidth required. finite load (un saturated) source/queue with: B ρ L/X, ρ 1 User Perception Rigid Traffic (UDP) Video/audio streaming Voice over IP Good Bad Minimum Bandwidth B Multimedia User Perception Elastic Traffic (TCP) Better Web (HTTP) FTP, e mail,... Worst B

8 The DCF/EDCA model

9 9 The scenario MN B=200 Kbps MN B=100 Kbps MN MN1 MN B=120 Kbps AP B=450 Kbps Be=max. av. MN B=max. av. MN B=max. av. MN MN2 B=max. av. MN B=80 Kbps MN and AP share the channel using a random access MAC protocol.

10 10 IEEE MAC Model (DCF) Each node is modelled by a M/M/1/K queue with packet arrival rate λ k and service time X. Assumption: Poisson arrivals and service time exponentially distributed. The model accounts for: queuing delays, packet losses. λ k X K EQ1:Queueing delay packet arrival (HOL) 1/λ MN1 packet arrival (queued) packet departure X1 1 Queue empty X1 1/λ MN2 packet departure 2 COLLISION t COLLISION t X2 backoff suspension

11 11 IEEE MAC Model (DCF) Each node is modelled by a M/M/1/K queue with packet arrival rate λ k and service time X. Assumption: Poisson arrivals and service time exponentially distributed. The model accounts for: queuing delays, packet losses. λ k X K EQ1:Queueing delay packet arrival (HOL) 1/λ MN1 packet arrival (queued) packet departure X1 1 Queue empty X1 1/λ MN2 packet departure 2 COLLISION t COLLISION t X2 backoff suspension

12 12 DCF EDCA EQ1:Queueing delay packet arrival (HOL) 1/λ MN1 packet arrival (queued) packet departure X1 1 1/λ MN2 2 packet departure t backoff suspension (coll.) Queue empty X1 empty slot MN3 COLLISION t COLLISION t X2 backoff suspension (succ.) BEB AIFS TXOP

13 13 IEEE e MAC Model (EDCA) : TXOP Multiple packet transmission each time a MN/AP gets the channel The overall network throughput increases (S=Data/Time [bps]) Less wasted time in random access. Packet arrival MN 1 Packet arrival (Queued) Packet arrival Packet arrival busy t busy Packet arrival (Queued) Packet arrival MN 2 Packet arrival (Queued) busy MN 2 MN 1 No TXOP Service Time (X) GAIN Packet arrival (Queued) busy busy TXOP t GAIN Service Time (X) t t

14 14 IEEE e MAC Model (EDCA) : TXOP Each user is modeled as an M/G[1,B]/1 queue (bulk service time queue) TXOP (Burst length): maximum consecutive number of frames transmitted at each successful attempt i=b Renewal Analysis j K Departure distribution. Average TXOP length Average TXOP duration Stationary (equil.) distribution. Blocking Probability. Queue occuppation. Queuing delay.

15 15 IEEE e MAC Model (EDCA): AIFS AIFS MN1 t Extra blocked slots due to AIFS Aggregate load from other nodes MN1 Service Time Transmission Prob. t

16 IEEE e MAC Model (EDCA): EB, p Expected BackOFF slots Cond. Collision Probability Prioritizing a flow: TXOP AIFS CWmin This model is able to capture the impact of the other flows with different MAC parameters. 16

17 17 Wireless Scenario Single Hop Ad Hoc network. No hidden terminals and ideal channel conditions. Channel access based on the DCF (BA and RTS/CTS). Two types of flows: streaming (UPD like) and elastic (TCP like). 2 1 i n 3 n 1

18 18 Results Single cell WLAN (ad hoc) Two types of flows: S1 (streaming ~ unsaturated) Bs = 100 Kbps L = 400 bytes E1 (elastic ~ saturated) Be (network dependant) L = 1500 bytes Default WLAN MAC parameters 0 S1 flows. Solution (1): AIFS (elastic flows, A=3) 8 S1 flows The TCP throughput is reduced. S1 saturation point

19 Model Validation Single cell WLAN (ad hoc) Two types of flows: S1 (streaming ~ unsaturated) Bs = 100 Kbps L = 400 bytes E1 (elastic ~ saturated) Be (network dependant) L = 1500 bytes Solution (2): TXOP S1 (=4) 7 S1 flows. CWmin,E1 (=64) 6 S1 flows. 19

20 20 Call Admission Control

21 Problem statement: Providing QoS Flow level Admission Control. Adaptive selection of MAC parameters. 21

22 22 Problem statement: Providing QoS Flow level Admission Control. Adaptive selection of MAC parameters. maximize performance (under the QoS) random access MAC (Performance)=f(MAC parameters, number of nodes, traffic profile each node) non-linear relation

23 23 The available bandwidth estimation problem Ideal Channel Sharing Random Access Channel Sharing Bandwidth Bandwidth Flow rejected Available bandwidth Bandwidth penalty Flow rejected Available bandwidth n... the bandwidth penalty due to the distributed access depends on the number of flows active and the traffic profile of each flow.... how to decide if a flow can be admited? We have to compute or to test if the new flow is possible. n

24 Multi rate Problems on VoIP capacity 24

25 25 Multirate wireless 11Mbps 11Mbps AP 11Mbps 11Mbps etc

26 26 Multirate wireless 1Mbps 11Mbps AP 11Mbps 11Mbps etc

27 27 Problem Statement While total cell capacity is around 12 calls when all of them transmit at 11Mbps this capacity falls with any change from fast to slow flows Distribution of VoIP flows in a cell (G.711 codec)

28 Multihop (Mesh) Networks VoIP capacity 28

29 29 VoIP Capacity in a mesh network VoIP calls (G.279) MP MP MP MP Single channel 2 Mbps Channel Contention (MAC) hidden nodes D. Niculescu. S. Ganguly, K. Kim, R. Izmailov; Performance of VoIP in a Wireless Mesh Network. IEEE Infocom 2006, Barcelona, Spain.

30 30 VoIP Capacity in a mesh network VoIP calls A MP MP MP MP Multiple channels (Two channels) Channel Contention (MAC) hidden nodes VoIP calls B MP MP MP MP NO Channel Contention (MAC) NO hidden nodes Multiple channels (Three channels)

31 VoIP Capacity in a mesh network D. Niculescu. S. Ganguly, K. Kim, R. Izmailov; Performance of VoIP in a Wireless Mesh Network. IEEE Infocom 2006, Barcelona, Spain. 31