1. Call Admission Control in IEEE 802.11 Wireless Networks using QP-CAT Sangho Shin Henning Schulzrinne Department of Computer Science Columbia University

2. 2 Call Admission Control (CAC) in IEEE 802.11 Wireless Networks QoS WIFI

3. 3 Call Admission Control (CAC) in IEEE 802.11 Wireless Networks WIFI QoS CAC

4. 4 Framework of CAC IEEE 802.11e Admission Control ADDTS Request Category TSpec ADDTS Response Category TSpec Status ? Min/Max MSDU Min/Max Service Interval Min/Avg/Max Data Rate WIFI

5. 5 Outline CAC in 802.11 Wireless Networks Related work QP-CAT Simulation results Experimental results Extension of QP-CAT Conclusion

6. 6 CAC in 802.11 Wireless Networks Problems Difficult to estimate QoS of VoIP traffic from the channel status Difficult to predict the impact of new VoIP calls Keys Accurate metric for QoS Need to represent delay not throughput Prediction algorithm Need to accurately predict the impact of new calls on QoS of existing calls

7. 7 Related work Model based Build a theoretical model Compute available bandwidth or delay Monitoring based Monitor the current transmissions Compute a metric (channel usage ratio etc.) Probing based Metric: delay and packet loss Used for wired networks Very accurate and simple Waste a certain amount of bandwidth Virtual Probing based QP-CAT

8. 8 QoS Metric in QP-CAT Metric: Queue size of the AP Strong correlation b/w the queue size of the AP and delay D=(Q+1)D T D=downlink delay D T =TX time of a VoIP frame

9. 9 QoS Metric in QP-CAT Estimation error

10. 10 Emulate new VoIP traffic Packets from a virtual new flow QP-CAT Algorithm (1/5) Basic flow of QP-CAT Compute Additional Transmission channel Actual packets Additional transmission Decrease the queue size Predict the future queue size + current packets additional packets

11. 11 QP-CAT Algorithm (2/5) Emulation of VoIP flows Two counters: DnCounter, UpCounter Follow the same behavior of new VoIP flows Increase the counters every packetization interval of the flows Decrement the counters alternatively 20ms time DnCounter++ UpCounter++ DnCounter++ UpCounter++ DnCounter++ UpCounter++ 20ms Example : 20ms packetization interval

12. 12 QP-CAT Algorithm (3/5) Computation of Additional Transmission

13. Handling T r Virtual Collision 13 QP-CAT Algorithm (4/5)

14. 14 QP-CAT Algorithm (5/5)

15. 15 QP-CAT 16 calls (actual) 17 calls + 1 virtual call (predicted by QP-CAT) 16 calls + 1 virtual call (predicted by QP-CAT) 17 calls (actual) 17th call is admitted 17 calls + 1 virtual call (predicted by QP-CAT) 16 calls + 1 virtual call (predicted by QP-CAT) 18th call starts 17 calls (actual) 18 calls (actual) Simulation results

16. 16 Experiments Linux, MadWifi, Atheros ORBIT test-bed in Rutgers University Experimental setup Ethernet-to-Wireless 11Mb/s data rate client clientsclientAPclient IEEE 802.11b

17. 17 QP-CAT Experimental results (64kb/s 20ms PI) 11Mb/s1 node - 2Mb/s 2 nodes - 2Mb/s 3 nodes - 2Mb/s

18. 18 Multiple execution of QP-CAT Parallel execution Need to test various types of VoIP traffic Run multiple QP-CAT using each type simultaneously Serial execution The longer we monitor, the better decision Takes time for accurate decision Run two QP-CAT serially

19. 19 QP-CATe QP-CAT with 802.11e Emulate the transmission during TXOP DDDTCP TXOP DDDTCP TcTc DDD TXOP CAC

20. 20 Conclusion QP-CAT uses the queue size of the AP as the metric for QoS of VoIP traffic QP-CAT can accurately predict the impact of new VoIP calls using CAT We can run QP-CAT in parallel or serially to handle multiple new VoIP flows QP-CAT can handle background traffic in 802.11e using QP-CATe

21. 21 Thank you

22. 22 QP-CAT Algorithm (4/8) Computation of Additional Transmission T c = T c2 + T r - T DIFS T r > T b T r < T b 12 TrTr 12 TrTr 12 TrTr T c2

23. 23 802.11 Frame Transmission DIFS Data SIFS ACK DIFS Data SIFS ACK Node A Node B Defer

24. 24 QP-CAT Algorithm (5/8) Handling T r : T r > T b

25. 25 QP-CAT Algorithm (6/8) Handling T r : T r < T b

26. 26 QP-CAT Algorithm (7/8) Virtual collision

27. 27 Implementation Environment Linux, MadWifi driver, Atheros chipset Monitoring Atheros chipset notifies RX timestamp in microsecond and TX timestamp in millisecond Additional wireless card as monitor mode at the AP Computing T C T C = RX 2 – RX 1 - T T RX 1 RX 2 TCTCT UpDn RXTX TCTC

28. 28 Related work (3/3) Comparison ApproachesMetricAssumptionAdapts to channel Waste of BWExtensibility Theoretical approaches CW/TXOP Computed bandwidth Saturated channel NoLowGood CUE/CBRCUE CBR Max CU/CB (Measured in advance) No (Fixed Max CU) Middle (Reserved BW for collisions) Good Actual Probing Delay packet loss NoYesHigh (Probing flow) Bad QP-CATQueue size of the AP NoYesLowGood