A Generic Maximum Delay Model of a Packet Switch

Abstract

There has been a strong trend away from shared medium (in the most recent case, the use of hubs) in local area networks in favor of switched local area networks. The need for deterministic guarantees on delays when designing switched local area networks has also been recognized by many researchers as these delays are useful engineering quantities. This is because, if the maximum delay between two nodes of a network is not known, it is impossible to provide a deterministic guarantee of worst case response time of packets' flows. In this paper, we describe a maximum delay model of a packet switch that can be used for designing maximum end-to-end delays packet switched networks. The packet switch model was obtained by using elementary components such as receive buffers, constant delay lines, multiplexer, first-in-firstout (FIFO) queue. The maximum delay value of the packet switch model was computed from an appropriate aggregation of the maximum delay values for the concatenated network elements. Comparison of the maximum packet delay value of the model with two other values obtained from literature showed that the model is better and much more realistic.

Key words: Packet Switch Model, Maximum End-To-End Delay, Network