Stochastic modeling of Scouting Switching for adaptively-routed mesh networks

摘要

Talking out the network issues, the switching techniques specify the connection activities performed by the switching elements when a message is received at the input port. Traditional switching mechanisms such as Wormhole Switching (WS) realize high performance, but prone to deadlock in the vicinity of faults. While some techniques such as adaptive routing can alleviate the problem, it cannot solve the problem by itself. This has motivated the development of different switching techniques. The Scouting Switching (SS) has been suggested as an efficient switching mechanism for reconciling the confliction demands on communication performance and fault-tolerance in interconnection networks. Although SS has been around for years and it can greatly benefit from adaptive routing as it reduces blocking in the network, there has been hardly any attempt to provide an analytical model for SS when fully adaptive routing along with virtual channels is used. Besides, mesh is one of the most desirable topologies regarding to characteristics which can offer very good scalability. In an effort to fill this gap, this paper proposes the first analytical model for 2-D mesh networks using SS augmented with virtual channels. Experimental results show that this model is able to predict message latency with a good degree of accuracy.