Transport Assistants to Enhance TCP Connections: Investigating the Placement Problem

Abstract

As of today, transport control protocol (TCP) remains the de-facto transport protocol in the Internet. However, TCP may incur high delays, especially when retransmitting lost packets as they have to be retransmitted only by the source and after a timeout that is roughly equal to a round trip time. To reduce such delay, recent work proposed to deploy a special network function, called transport assistant (TA), that is able to detect and retransmit lost TCP packets from inside the network rather than the source and thereby reduces retransmission delays. Unfortunately, there is no study on the impact of the placement of the TA on its performance benefits in terms of packet delivery delay. In this paper, we focus on the TA placement problem. We discuss the trade-offs and parameters to be considered to select the best placement for the TA. We first mathematically model the TCP packet delivery delay, that is, the time needed to deliver TCP packets, when the TA is deployed. We also formulate, as an integer linear program (ILP), the problem of placing multiple TAs in order to reduce TCP packet delivery delays while minimizing their deployment costs. We consider use-cases, one where a TA could handle a single flow and another where a TA could handle multiple flows. We then propose two heuristics to solve the problem with minimal execution time. Through experiments, we demonstrated that the deployment of TAs could reduce TCP packet delivery delays by up to 30% and could be leveraged to guide routing and load balancing. Moreover, we show that using the proposed heuristics for placing TAs could lead to performance that is close to the optimal solutions obtained with the ILP but with lower execution time.