Network-based parallel computing over ATM using improved SSCOP protocol

Abstract

In this paper, we address the problem of supporting network-based parallel computing, usually known in the literature as High Performance Distributed Computing (HPDC), over ATM networks. For this purpose, we consider a logically separate subnetwork for HPDC applications. After presenting an architectural reference model for the HPDC subnetwork and distinguishing which functions should be installed over the ATM network in order to satisfy the needs of HPDC applications, we suggest two different modifications of the loss recovery mechanism of the Service-Specific Connection Oriented Protocol (SSCOP) that aim at optimizing communications by taking advantage of both the special properties of HPDC traffic and the cell-based nature of ATM. The performance of these mechanisms is evaluated and compared with that achieved by the standard mechanism of the SSCOP protocol. The results show that, when the ATM network experiences high load and the HPDC applications make an intensive use of arrays, SSCOP including the new retransmission mechanisms becomes more robust than standard SSCOP, and provides low latency and efficient cell loss recovery. From the results, we conclude that the introduction of cell-based retransmission mechanisms does contribute to enhance the performance of ATM network-based parallel computing.