CCStarve

The Internet has grown from a small research network to a global information infrastructure that has assumed tremendous importance in our everyday life. The success of the Internet can partly be attributed to the congestion control algorithm in the Transmission Control Protocol (TCP). However, TCP is now showing signs of ageing and many of the assumptions made during its inception no longer hold true. With the incorporation of high speed, long distance links on the Internet and the deployment of access networks such as 802.11 LANs and multi-hop wireless mesh networks, TCP performance issues are becoming increasingly problematic.

TCP was designed to suit an environment where the Bandwidth-Delay Product (BDP) or ”pipe size” of the paths was typically less than ten packets and any packet loss inside the network was assumed to be due to overflow of router buffers. However, today, BDP of many Internet paths is orders of magnitude larger and with the proliferation of wireless LANs and multi-hop wireless networks, assuming congestion to be the only source of packet loss can be very constraining because (a) bit errors, hand-offs, multi-path fading etc., account for a significant proportion of lost packets and (b) responses to different kinds of losses need to be different.

This project aims to investigate TCP performance in dense Access Point (AP) deployments and simple multi-hop wireless mesh networks. Specifically, we plan to quantify starvation issues in such scenarios that have been reported in prior studies using simulations. We then aim to develop schemes for addressing these issues in such networks.

People

· Daniel Mosse

· Ihsan Ayyub Qazi

· Taieb Znati

Evaluation Testbed

We are using Emulab for conducting our experiments.