Assessment of network protocols and software tools for distributed computing

Abstract

The performance of distributed supercomputing computing environments are mainly dependent on three factors: distributed programming tools, computing nodes, and LANs employed. In this paper, we analyze the performance of all these factors experimentally and analytically. The distributed programming tools that we employed are PVM and Express. The computing nodes that we employed are SUN and HP workstations, and the LANs that we considered are an Ethernet and FDDI networks. Extensive timing experiments, including one-to-one communications, exchange operations, and broadcast operations, have been performed and analyzed. Moreover, analytic models have been developed to analyze the behavior of the network protocols employed by the LAN-based platforms as well as to estimate the communication overhead for the computing software tools.