As broadband connection capacities increase, residential network users can experience performance degradation when multiple applications run concurrently over bottlenecked wireless access points (APs). This paper presents Classification and Treatment iN an AP (CATNAP), a plug-and-play solution for an AP that automatically classifies and treats flows passing through without any user intervention. CATNAP classifies application flows across three dimensions based on their response to flow traffic, bitrate greediness and interactivity. Then, based on the classification, CATNAP treats flows by rate limiting/preserving, adjusting packet time in queue or dropping packets. CATNAP is implemented in NS-2 and evaluated against DropTail and Strict Priority Queuing under various network and traffic conditions. Analysis of the results show CATNAP provides better QoS support for VoIP, games and streaming video applications.

Download:

• Paper: pdf
• Slides: pptx

See also:

• Feng Li. Treatment-Based Classification in Residential Wireless Access Points, Ph.D. Thesis, Computer Science Department, Worcester Polytechnic Institute, Spring 2014. (Committee: Professor Mark Claypool (WPI), Professor Robert Kinicki (WPI), Professor Craig Wills (WPI), and Dr. Tim Strayer (BBN Technologies)) Online at: http://www.cs.wpi.edu/~claypool/phd/lif/

• Mark Claypool, Robert Kinicki and Craig Wills. Treatment-Based Traffic Signatures, Invited presentation at IMRG Workshop on Application Classification and Identification (WACI), Cambridge, Massachusetts, USA, October 2007. Online at: http://www.cs.wpi.edu/~claypool/papers/cube/