Continuous time performance analysis for a hierarchical routing algorithm in Manhattan street networks

The Manhattan street network (MSN) is a directed network with a link topology that resembles the one-way system of streets and avenues in midtown Manhattan. It is designed for packet communications in a local or metropolitan area. In order to improve on the delay-throughput performance of the random routing algorithm yet keep its properties of simplicity and flexibility, a hierarchical deflection routing algorithm is proposed. We conduct theoretical analyses of this routing algorithm in continuous time for both the store and forward model and the virtual cut-through model, and compare the results for both models.<>