TCP Pegas: A PSO-based improvement over TCP Vegas

TCP Vegas is a source algorithm that offers relatively rich performance in the Internet congestion control. But Vegas has some problems which have serious impacts on its performance. Rerouting is one of these problems. When route of a connection changes and round trip time increases, Vegas misinterprets it as the result of the network congestion and consequently decreases its own sending rate. As another important problem, when a flow joints to the network later than other flows and faces with congested queues, it wrongly considers the measured round trip time as its initial Base(RTT). It means that while other flows decrease their sending rates due to existing congestion, this flow does not sense the congestion and hence unfairly increases its sending rate. These problems mainly have roots in the Vegas estimation procedure of the propagation delay i.e. Base(RTT). In this paper we propose a novel algorithm, named Pegas, in which particle swarm optimization technique is used to dynamic estimation of Base(RTT). Simulation results show that Pegas solves the rerouting and unfairness problems and remarkably enhances Vegas performance in terms of dropped packets, bottleneck utilization, and fairness. (C) 2015 Elsevier B.V. All rights reserved.