Modeling the Transient Rate Behavior of Bandwidth Sharing as a Hybrid Control System

This paper uses hybrid control to model a problem of computer network systems, the dynamic behavior of bandwidth sharing among competing TCP traffic. It has been well known in the computer network community that well-behaved (TCP-friendly) congestion control mechanisms are crucial to the robustness of the Internet. Congestion control determines the transmission rate for each flow. Right now, most TCP-friendly research focuses only on the average throughput behavior without considering how the data is sent out in the short-term (e.g. bursty or smooth). However, recent experimental results show that short-term rate adjustments can change the bandwidth sharing result. Therefore, it is important to study the dynamic behavior of bandwidth sharing, and its impact on the long-term average throughput behavior. While existing network models and simulators for TCP-friendly control work well in general, they can not be used for theoretical proofs of the stability of the bandwidth-sharing behavior by competing TCP-friendly congestion controls. We want models that can capture the system’s dynamic behavior to help us understand and predict the impact of short-term rate adjustments. In this paper, we present a hybrid state-space-based model for the bandwidth sharing among TCP-friendly flows. We explain the hybrid nature of the system, which has a distributed control algorithm and is event-driven. The model helps us to understand the system’s behavior in general and we use it to prove the system’s stability under certain assumptions about packet loss. We also describe simulations and experiments to explore the cases in which these assumptions do not hold. This paper presents some preliminary results.