Global stability of Vegas-like TCP flow

Abstract

A TCP Vegas flow adapts its sending rate to maintain a constant backlog in its path. The stability of nonlinear adaptation has been analyzed based on linearization and only accounted for a small signal. We extend the error model of TCP-like flow to a state-dependent coefficient form with nonlinear state feedback. The nonlinear feedback is here approximated by a saturation function. Using a quadratic Lyapunov function approach, we find a domain of attraction to show that the unique equilibrium point of the system is asymptotically stable in the domain.

Files

Additional details