Impact of interference and capture effects in 802.11 wireless networks on TCP

As broadband wireless channels become common, the performance of TCP over end-to-end paths containing such links is important. TCP SACK suffers substantially when residual packet error rates increase beyond a value of about 1%--5%(especially for longer RTTs). Recently we have proposed improvements to TCP (called LT-TCP) to make TCP loss-tolerant in heavy and bursty erasure environments. However, real world wireless systems do not just present bursty randomloss patterns to the transport layer. The PHY, MAC and transport layers all respond to errors, interacting in myriad ways. In this paper, we focus on one underlying source of packet erasure (non-congestion loss), namely interference in 802.11 environments (from Bluetooth and co-channel interferers), and the resulting interaction between the MAC and transport layer mechanisms. MAC layer mechanisms cannot fully mitigate the interference problem and tend to misinterpret interference as noise and aggressively respond with techniques like rateadaptation. Such aggressive responses lead to poor scheduling performance at the MAC layer (e.g., well-known unfairness and capture effects) and limit mitigation opportunities at the transport layer. We argue that aggressive rate adaptation is undesirable in these situations and show how a combination of reconfiguration of MAC layer mitigation options and increased buffering leads to significantly improved end-to-end performance.

[1]  K. K. Ramakrishnan,et al.  LT-TCP: End-to-End Framework to Improve TCP Performance over Networks with Lossy Channels , 2005, IWQoS.

[2]  Kimberly C. Claffy,et al.  Measurement and visualization of internet connectivity and performance , 2001 .

[3]  Edward W. Knightly,et al.  Opportunistic media access for multirate ad hoc networks , 2002, MobiCom '02.

[4]  Haiyun Luo,et al.  The impact of multihop wireless channel on TCP throughput and loss , 2003, IEEE INFOCOM 2003. Twenty-second Annual Joint Conference of the IEEE Computer and Communications Societies (IEEE Cat. No.03CH37428).

[5]  Thierry Turletti,et al.  IEEE 802.11 rate adaptation: a practical approach , 2004, MSWiM '04.

[6]  David L. Black,et al.  The Addition of Explicit Congestion Notification (ECN) to IP , 2001, RFC.

[7]  Luigi Rizzo,et al.  On the feasibility of software FEC , 1997 .

[8]  Gunnar Karlsson,et al.  Adaptive end-to-end FEC for improving TCP performance over wireless links , 2004, 2004 IEEE International Conference on Communications (IEEE Cat. No.04CH37577).

[9]  Paramvir Bahl,et al.  A rate-adaptive MAC protocol for multi-Hop wireless networks , 2001, MobiCom '01.

[10]  Christoph Lindemann,et al.  TCP with adaptive pacing for multihop wireless networks , 2005, MobiHoc '05.

[11]  Haiyun Luo,et al.  The impact of multihop wireless channel on TCP performance , 2005, IEEE Transactions on Mobile Computing.

[12]  Nada Golmie,et al.  Interference in the 2.4 GHz ISM band: impact on the Bluetooth access control performance , 2001, ICC 2001. IEEE International Conference on Communications. Conference Record (Cat. No.01CH37240).

[13]  Ren Wang,et al.  TCP westwood: Bandwidth estimation for enhanced transport over wireless links , 2001, MobiCom '01.

[14]  Rajesh Krishnan,et al.  Explicit transport error notification (ETEN) for error-prone wireless and satellite networks , 2004, Comput. Networks.

[15]  Nada Golmie,et al.  Interference of bluetooth and IEEE 802.11: simulation modeling and performance evaluation , 2001, MSWIM '01.

[16]  Ramesh R. Rao,et al.  Performance of IEEE 802.11 WLANs in a Bluetooth environment , 2000, 2000 IEEE Wireless Communications and Networking Conference. Conference Record (Cat. No.00TH8540).

[17]  Nada Golmie,et al.  Interference Evaluation of Bluetooth and IEEE 802.11b Systems , 2003, Wirel. Networks.