An Argument for Increasing TCP's Initial Congestion Window

In this paper, the authors propose to increase TCP's initial congestion window to at least ten segments (about 15KB). Through large-scale Internet experiments, the authors quantify the latency benefits and costs of using a larger window, as functions of network bandwidth, round-trip time (RTT), band width delay product (BDP), and nature of applications. The authors show that the average latency of HTTP responses improved by approximately 10% with the largest benefits being demonstrated in high RTT and BDP networks.