Invited Talk by Ms.Mukulika Maity on TCP Download Performance in Dense WiFi Scenarios: Analysis and Solution
How does a dense WiFi network perform, specifically for the common case of TCP download? While the empirical answer to this question is 'poor', analysis and experimentation in prior work has indicated that TCP clocks itself quite well, avoiding contention-driven WiFi overload in dense settings.
This talk focuses on measurements from a real-life use of WiFi in a dense scenario: a classroom where several students use the network to download quizzes and instruction material. We find that the TCP download performance is poor, contrary to that suggested by prior work. Through careful analysis, we explain the channel contention and complex interaction of various phenomena which leads to this poor performance. To improve performance, we propose a solution WiFiRR, which picks only a subset of clients "active" during any instant, and varies this set periodically in a round-robin fashion over all clients to ensure that no client starves. We evaluate WiFiRR extensively in simulation and in real settings. By reducing the number of contending nodes at any point of time, WiFiRR improves the download time of large TCP flows upto 3.5× of our classroom scenario.
Mukulika Maity received her B.E. in Computer Science and Engineering from Bengal Engineering and Science University, Shibpur in 2010. She joined Computer Science and Engineering Dept. of Indian Institute of Technology, Bombay in 2010 to pursue M.Tech. In 2012, she joined dual degree program (M.Tech. + PhD). Her PhD topic is health diagnosis and congestion mitigation of WiFi networks. She works with Prof. Bhaskaran Raman and Prof. Mythili Vutukuru. Her research interests are broadly in the area of wireless networks, mobile computing.