Performance modelling of adaptive routing in hypercubic networks under non-uniform and batch arrival traffic

Traffic loads have a significant impact on the performance of routing algorithms. Many analytical models for adaptive routing in interconnection networks have been reported. However, most existing studies are based on the assumption that the arrivals of traffic follow a non-bursty Poisson process and the message destinations are uniformly distributed over the network. With the aim of obtaining a deep understanding of network performance under more realistic working conditions, this study develops an analytical performance model for adaptive-routed hypercubic networks under hot-spot and batch arrival traffic. This model adopts the Compound Poisson Process (CPP) to capture the properties of the batch arrival traffic. Extensive simulation experiments are conducted to validate the accuracy of the analytical model.