A new hierarchical interconnection network and its performance analysis

We propose a new hierarchical interconnection network (HIN) which can be used to build large cluster-based systems. The proposed HIN offers a number of advantages: increased fault tolerance; better performance; and, ability to be used in a cache based system. An analytical model has been developed to measure the performance of the proposed network. The analytical model has been developed based on the assumption that “the requests which are not accepted during a cycle are blocked and resubmitted during the next memory cycle”, in contrast to the commonly made assumption that “the requests which are not accepted during a cycle are discarded”. The performance of the proposed HIN is measured in terms of memory bandwidth of the system. For the same set of system parameters and the same set of assumptions for the analytical models, we compared the performance of our system with the system presented by Mahgoub (see IEEE Transactions on Parallel and Distributed Systems, vol. 3, no. 2, p129-38, 1992)