DESIGN GUIDELINES FOR DISTRIBUTED COMPUTING SYSTEMS

The existence of different computing platforms (such as workstations, minicomputers, mainframes, and supercomputers) has led to an increased interest in distributed computing systems. Significant increase in performance, reliability, and fault tolerance can be obtained because they are inherently redundant. Furthermore, distributed computing environments provide cost-effective means for resource sharing and extensibility. The design and engineering of distributed computing environments are difficult and more complex than the design and engineering of centralized computing systems because of heterogeneously dispersed components and component interactions. This paper describes the guidelines that help to systematically address the various design issues of distributed systems and to make practical engineering decisions. The design phases of distributed systems can be broadly identified as follows: (1) requirements specifications; (2) conceptual design; (3) functional architecture; (4) design analysis and verification; and (5) implementation. In this paper, focus is placed on the requirements specification phase and several alternatives available to examine the requirements are investigated. These requirements are to be studied with respect to hardware/software, configuration, functions and services provided to users, performance, reliability, and cost.