Performance index - A metric to analyze and evaluate performance of hypercube interconnection networks

Abstract

A Multiprocessor is a computing machine which has at least two central processors, to process the tasks simultaneously. An Interconnection network that links multiple processors greatly influences the performing capability of the entire system. Various network scales to evaluate interconnection networks are: degree, speed, node coverage, reliability, scalability, diameter, connectivity, throughput, packet loss, and network cost etc. Interconnection Network like a Hypercube can be considered as an graph with undirected edges, where a vertex indicates a processor and an undirected edge denotes a communication medium among the processors. Some of the variants of Hypercube Interconnection Networks are Hypercube Network, Folded Hypercube Network, Multiple Reduced Hypercube, Multiply Twisted Cube, Recursive Circulant, Exchanged Crossed Cube Network, Half Hypercube Network etc. The vital purpose of this paper is to investigate different variants of hypercube interconnection networks and to analyze their properties to summarize the differences in their performance. It is also discussed how to analyze and evaluate the performance of Hypercube Networks and identify the changes in Performance Index based on the variations in their properties.