Compound TCP: A scalable and TCP-friendly congestion control for high-speed networks

Abstract

Many applications require fast data transfer over high speed and long distance networks. However, standard TCP fails to fully utilize the network capacity in high-speed and long distance networks due to its conservative congestion control (CC) algorithm. Some works have been proposed to improve the connection’s throughput by adopting more aggressive loss-based CC algorithms, which may severely decrease the throughput of regular TCP flows sharing the network path. On the other hand, pure delay-based approaches may not work well if they compete with loss-based flows. In this paper, we propose a novel Compound TCP (CTCP) approach, which is a synergy of delay-based and loss-based approach. More specifically, we add a scalable delay-based component into the standard TCP Reno congestion avoidance algorithm (a.k.a., the loss-based component). The sending rate of CTCP is controlled by both components. This new delay-based component can rapidly increase sending rate when the network path is under utilized, but gracefully retreat in a busy network when a bottleneck queue is built. Augmented with this delay-based component, CTCP provides very good bandwidth scalability and at the same time achieves good TCP-fairness. We conduct extensive packet level simulations and test our CTCP implementation on the Windows platform over a production high-speed network link in the Microsoft intranet. Our simulation and experiments results verify the properties of CTCP.