Switching and scheduling in ATM and IP over ATM networks.

Abstract

The key to multi-service platform performance with QoS support is to provide appropriate switching and scheduling mechanisms at various contention points in a network. In this thesis, we focus on the two key elements in any traffic scheme: switching and scheduling. We introduce innovative switching architecture and scheduling algorithms designed to keep them in line with network operations. In ATM networks, we propose per-VC queueing architecture and per-VC scheduling. In IP networks, we propose per-flow queueing in access routers and VC-merge in core routers to address the scalability issue. We also introduce per-flow scheduling in tandem to our switching architecture. We conduct performance analysis to evaluate the performance of our proposed switching and scheduling schemes. Simulations are used to validate our analytical results.