Congestion Control: Model and Algorithms
Yu‐Ping Tian · 2012
This chapter introduces basic notions, algorithms and system models of end-to-end congestion control. The congestion control problem is modeled as an optimization problem of resource allocation in the network under link capacity constraints, and the congestion control algorithms are treated as distributed real-time gradient-descent algorithms of seeking for the optimal solution of the resource allocation problem. The chapter shows the primal algorithm has a dynamical law for adjusting source rate and a static law for generating link price, and such an algorithm properly describes currently used congestion avoidance protocols for TCP rate control at sources. Similarly, the dual algorithm has a dynamical law for adjusting link price and a static law for generating source rate, and such an algorithm can be used for the AQM mechanism at links. Controlled Vocabulary Terms networked control systems; TCPIP; telecommunication congestion control