Admission Control for IEEE 802.11e Wireless LANs

1. Admission Control for IEEE 802.11e Wireless LANs Student: Conroy Smith Supervisor: Neco Ventura Department of Electrical Engineering University of Cape Town

2. Overview • Introduction • IEEE 802.11e QoS Enhancement • Proposed Problem • Proposed Admission Control solution • Performance Evaluation and Preliminary Results • Conclusions • Future Work

3. Introduction • Wireless LANs are expected to have a major impact on people’s daily life styles • Provides Cheaper Internet Connectivity • Relatively high throughput • Ease of Implementation • Being endowed with roaming capabilities • Voice enabled devices are now being equipped with WiFi capabilities • This allows WiFi to compete directly with 3G Cellular Networks The Internet

4. Introduction • Original 802.11 Standard Lacks QoS Support • WLANS was traditionally a “Best Effort” Wireless Access • IEEE 802.11e provides a QoS enhancement • Modifications made to MAC layer • Provides service differentiation • QoS can only be achieved at when the network load is not too heavy • Channel overloading decreases network throughput

5. IEEE 802.11e QoS enhancement • IEEE 802.11e Specifies are new Hybrid Co-ordination Function (HCF) • The HCF Specifies 2 Access modes: • HCF Controlled Channel Access (HCCA) • Enhanced Distributed Channel Access (EDCA) MAC Layer Modification of the IEEE 802.11e enhancement

6. IEEE 802.11e EDCA • The EDCA allow service differentiation, by supporting 8 different priorities • further mapped to 4 Access Classes (ACs) • Each AC behaves as a single enhanced DCF contending entity with dedicated queues

7. IEEE 802.11e EDCA • Differentiation achieved by: • AIFS • CWmin • CWmax • TXOP Queuing Architecture of EDCA

8. IEEE 802.11e EDCA • Contention in the EDCA:

9. Problem of overloading 802.11 channel in the EDCA • Traffic congestion and can lead to severe overall network degradation. • New flows will may not be able to achieve their QoS • They may also degrade the QoS of other admitted flows • The need for Admission control has become apparent

10. Proposed Admission Control For IEEE 802.11e • A measurement/model-based Admission control • WLAN channels are modeled using the a modified Bianchi Model [4] • This is used to Bandwidth Estimations for each Virtual Station • Each AC queue act as a Virtual Station • Queue utilization and collision statistics are measured and used by the model • Uses TSPEC to negotiate Admission control • Flows must state their throughput requirements

11. Admission Negotiation Typical TSPEC Negotiation

12. Admission Control Decisions • A new request will demand more Throughput for its “Virtual Station” • New flows will be Accepted ONLY if: • Achievable Bandwidth (Si) ≥ Requested Bandwidth For All Virtual Stations • Achievable Bandwidth: P(C) - Probability of a collision in a slot P(I) - Probability of an idle slot P(S) - Probability of a successful Tx in a slot P(S|VS=i) - Probability of a successful Tx on Virtual Station i

13. Admission Control Decisions: Calculating the Probabilities

14. Calculating Transmission Probability for each Virtual Station - Measured Collision Probability - Percentage of time that the Virtual station is considered to be Saturated • A modified Bianchi model is used to calculate the Transmission Probabilities for each Virtual Station

15. Accuracy of Bandwidth Estimations • Bandwidth Estimation Framework Integrated in the NS-2 Contributed Model from [5] • Simulation Consists of: • 1 AP • 6 Stations, 3 have unlimited data to send and 3 are unsaturated • All Stations transmit at 18 Mbps (802.11a)

16. Performance Evaluation of proposed admission control solution Accuracy of the Bandwidth Estimations

17. Performance Evaluation: Simulation Set up • At time t = 3 sec, each station has a TCP session and a voice and video flow • A new voice and video requests are added every 2 seconds • Voice flows: 64 Kbps • Video flows: 750 Kbps

18. Preliminary Results Performance without Admission Control (TCP session + Voice flow + Video flow)

19. Preliminary Results Performance with Admission Control (TCP session + Voice flow + Video flow)

20. Preliminary Results Performance without Admission Control (Total Throughput) Flow 16 admitted

21. Preliminary Results Performance with Admission Control (Total Throughput) Flow 16 rejected • Admitted Flows: • 9 Voice • 7 Video Flow 18 rejected

22. Conclusions • IEEE 802.11e WLANs provides QoS Support • EDCA Allows differentiation of 4 ACs Differentiated by, CW, TXOP and AIFS • Channel overloading can lead to severe degradation of QoS for EDCA flows • A measurement aided model-based admission control solution is proposed to protect QoS for EDCA flows • The accuracy of the bandwidth estimation indicates that effective admission control decisions can be made

23. Future Work • Extend the Admission Control scheme to be compatible with TXOP Bursting and RTS/CTS handshake mechanisms • Investigate whether acceptable delays are achieved

24. References [1] Y. Xiao and H. Li, “Evaluation of Distributed Admission Control for the IEEE 802.11e EDCA“, IEEE Communications Magazine, vol. 42, no. 9, pp. S20–S24 2004 [2] D. Gu and J. Zhang, “A New Measurement-based Admission Control Method for IEEE 802.11 Wireless Local Area Networks”, Mitsubishi Elec. Research Lab, Tech. rep. TR-2003-122, Oct. 2003. [3] D. Pong and T. Moors, “Call Admission Control for IEEE 802.11 ContentionAccess Mechanism”, Proc. IEEE GLOBECOM’03, vol. 1, San Francisco, CA, Dec. 2003, pp. 174–78. [4] G. Bianchi, “Performance Analysis of the IEEE 802.11 Distributed Coordination Function”, IEEE JSAC, 18(3): 535-47, Mar. 2000. [5] http://yans.inria.fr/ns-2-80211/ [6] H. Wu et. al. “IEEE 802.11e Enhanced Distributed Channel Access (EDCA) Throughput Analysis” [7] Z Kong et. al. “Performance Analysis of IEEE 802.11e Contention-Based Channel Access” [8] J. F. Robinson and T. S. Randhawa, “Saturation Throughput Analysis of IEEE 802.11e Enhanced Distributed Coordination Function”

25. Questions ?