Analytical modeling of a fat-tree network with buffered switches

Abstract

A performance analysis model of a fat-tree network with the multiple buffered crossbar switches is proposed and examined. The buffered switch technique is well known to solve the data collision problem in the switch network. The proposed evaluation model is developed by formulating the transfer pattern of data packets in a switch. The performance of the multiple-buffered a/spl times/b crossbar switch is then analyzed. The steady state probability concept is used to simplify the analyzing processes. The proposed model takes simple and primitive switch networks, i.e., no flow control and drop packet, to demonstrate analysis procedures clearly. It, however, can not only be applied to any other complicated modern switch networks that have intelligent flow control but also estimate the performance of any size networks with multiple-buffered switches. To validate the proposed analysis model, the simulation is carried out on the various sizes of fat-tree networks that use the multiple buffered crossbar switches.