This report examines potential stability problems associated with a model based apporach to TCP-friendly flow control for non-TCP traffic. In specific, such an approach involves using a TCP-friendly formula that estimates the throughput of a TCP session with the same end-to-end traffic characteristics as the non-TCP connection under consideration. The inputs to this formula include the round trip time, the timeout value, and the packet loss fraction of the connection. This paper shows that estimating the loss fraction per transmitted packet highly depends on the current transmission rate of the connection as well as the actual loss fraction of the path. Thus the estimated loss fraction can contain errors which result in inaccurate estimation of the corresponding TCP throughput. This inaccuracy can push the transmission rate of the non-TCP connection away from the fair share of the bottleneck bandwidth on the end-to-end path, so that under steady state, the connection ends up receiving either over-allocation or under-allocation of bandwidth.
Cited By
- Widmer J, Boutremans C and Boudec J (2004). End-to-end congestion control for TCP-friendly flows with variable packet size, ACM SIGCOMM Computer Communication Review, 34:2, (137-151), Online publication date: 1-Apr-2004.
- Widmer J, Mauve M and Damm J Probabilistic congestion control for non-adaptable flows Proceedings of the 12th international workshop on Network and operating systems support for digital audio and video, (13-21)
- Jin T, Sheng X and Wu W The Effect on the Inter-Fairness of TCP and TFRC by the phase of TCP Traffics Proceedings of the 2001 International Conference on Computer Networks and Mobile Computing (ICCNMC'01)
Recommendations
Network Tomography-Based Unresponsive Flow Detection and Control
FTDCS '03: Proceedings of the The Ninth IEEE Workshop on Future Trends of Distributed Computing SystemsTo avoid a congestion collapse, network flows should adjust their sending rates. Adaptive flows adjust the rate, while unresponsive flows do not respond to congestion and keep sending packets. Unresponsive flows waste resources by taking their share of ...
Improvement of BBRv2 Congestion Control Algorithm Based on Flow-aware ECN
Google proposed a new congestion control algorithm (CCA) based on bottleneck bandwidth and round-trip propagation time (BBR), which is considered to open a new era of congestion control. BBR creates a network path model by measuring the available ...