Stability of TCP Dynamics in Large Data Networks

We study the stability of the dynamics of a deterministic model of Transmission Control Protocol (TCP) traffic in a multiple flow, multiple router data network. In this model, flow rates increase continuously until network congestion causes them to decrease discontinuously. In computer simulations of small networks, trajectories appear to approach periodic orbits for most but not all parameter values. We prove that if all the data flows are coupled, periodic orbits must be exponentially attracting and thus persist under parameter changes, regardless of network size. Furthermore, we describe the model as a discontinuous but piecewise affine map and show that trajectories must either approach a periodic orbit or come arbitrarily close to map discontinuities.

[1]  V. Jacobson,et al.  Congestion avoidance and control , 1988, CCRV.

[2]  Douglas J. Leith,et al.  Partitioning and Invariance of AIMD Dynamics in Synchronised Networks. , 2006 .

[3]  QUTdN QeO,et al.  Random early detection gateways for congestion avoidance , 1993, TNET.

[4]  François Baccelli,et al.  AIMD, fairness and fractal scaling of TCP traffic , 2002, Proceedings.Twenty-First Annual Joint Conference of the IEEE Computer and Communications Societies.

[5]  R. La,et al.  Nonlinear instabilities in TCP-RED , 2004, Proceedings.Twenty-First Annual Joint Conference of the IEEE Computer and Communications Societies.

[6]  F. Baccelli,et al.  Interaction of TCP flows as billiards , 2003, IEEE INFOCOM 2003. Twenty-second Annual Joint Conference of the IEEE Computer and Communications Societies (IEEE Cat. No.03CH37428).

[7]  Sally Floyd,et al.  TCP and explicit congestion notification , 1994, CCRV.

[8]  Van Jacobson,et al.  Random early detection gateways for congestion avoidance , 1993, TNET.

[9]  Fabian R. Wirth,et al.  A positive systems model of TCP-like congestion control: asymptotic results , 2006, IEEE/ACM Transactions on Networking.