Why flow-completion time is the right metric for congestion control

Users typically want their flows to complete as quickly as possible. This makes Flow Completion Time (FCT) an im portant - arguably the most important - performance metric for the user. Yet research on congestion control focuses almost entirely on maximizing link throughput, utilization and fairness, which matter more to the operator than the user. In this paper we show that with typical Internet flow sizes, existing (TCP Reno) and newly proposed (XCP) congestion control algorithms make flows last much longer than necessary - often by one or two orders of magnitude. In contrast, we show how a new and practical algorithm - RCP (Rate Control Protocol) - enables flows to complete close to the minimum possible

[1]  Tom Kelly,et al.  Scalable TCP: improving performance in highspeed wide area networks , 2003, CCRV.

[2]  Andrew Odlyzko,et al.  Implications for the long distance network , 2003 .

[3]  David A. Patterson,et al.  Latency lags bandwith , 2004, CACM.

[4]  David A. Patterson,et al.  Latency Lags Bandwidth , 2005, ICCD.

[5]  Nick McKeown,et al.  Processor Sharing Flows in the Internet , 2005, IWQoS.

[6]  Linus Schrage,et al.  Letter to the Editor - A Proof of the Optimality of the Shortest Remaining Processing Time Discipline , 1968, Oper. Res..

[7]  McKeownNick,et al.  Why flow-completion time is the right metric for congestion control , 2006 .

[8]  Ronald W. Wolff,et al.  Stochastic Modeling and the Theory of Queues , 1989 .

[9]  Joseph Y.-T. Leung,et al.  Minimizing Mean Flow Time with Release Time Constraint , 1990, Theor. Comput. Sci..

[10]  Alexandre Proutière,et al.  Statistical bandwidth sharing: a study of congestion at flow level , 2001, SIGCOMM.

[11]  Mark Handley,et al.  Congestion control for high bandwidth-delay product networks , 2002, SIGCOMM.

[12]  Andrew M. Odlyzko.  The Internet and other networks: utilization rates and their implications , 2000, Inf. Econ. Policy.

[13]  Thomas Bonald,et al.  Statistical bandwidth sharing: a study of congestion at flow level , 2001, SIGCOMM.

[14]  Mischa Schwartz,et al.  ACM SIGCOMM computer communication review , 2001, CCRV.

[15]  Abhay Parekh,et al.  A generalized processor sharing approach to flow control in integrated services networks-the single node case , 1992, [Proceedings] IEEE INFOCOM '92: The Conference on Computer Communications.

[16]  Jan Karel Lenstra,et al.  Complexity of machine scheduling problems , 1975 .

[17]  Sally Floyd,et al.  HighSpeed TCP for Large Congestion Windows , 2003, RFC.

[18]  Abhay Parekh,et al.  A generalized processor sharing approach to flow control in integrated services networks: the single-node case , 1993, TNET.

[19]  Scott Shenker,et al.  Analysis and simulation of a fair queueing algorithm , 1989, SIGCOMM '89.