Date of Award

12-1993

Degree Name

Master of Science

Department

Computer Science

First Advisor

Dr. Ajay Gupta

Second Advisor

Dr. Thomas F. Piatkowski

Third Advisor

Dr. Elise deDoncker

Access Setting

Masters Thesis-Open Access

Abstract

A parallel and distributed processing environment can be defined as one where a set of workstations is configured in a certain topology (such as completely connected linear chain) to simulate the working of a particular parallel architecture. Such an environment provides an extremely useful means of experimenting with parallel algorithms, without the use of expensive dedicated parallel machines.

Several parallel and distributed processing environments exist, such as Parallel Virtual Machine, The Condor System and the Reactive Kernel System/ Cosmic Environment System. Each of these systems has some unique features and limitations. Other systems include p4, Hermes, Linda and Express.

This thesis describes the HPCS system and its three modes of operation with examples and briefly discusses future expansion potentials.

In this thesis, we describe the design, development and features of the Heterogeneous Parallel Computing System, a parallel and distributed processing environment, which allows program development given a limited set of resources and in a reasonable amount of time. It employs three modes of operation to suit the user's needs; the central-server mode and distributed-server mode are client-server models and the client-serve mode of HPCS is a serverless model. The HPCS system has been conceptually designed to overcome limitations posed by the well-known systems (e.g. PVM, RK/CE, Condor).

Share

COinS