Some Performance Studies of Distributed Systems

Abstract

A distributed system comprises of several software and hardware components that newlineforms a network of systems for communicating and coordinating the actions. In other newlinewords, a distributed computation constitutes a group of processes that cooperate to newlineachieve a common goal. The systems are distributed across a huge geographical region newlineand communicate with each other for executing the tasks. The different types of newlinedistributed systems include Distributed Computing Systems, Distributed Information newlineSystems and Distributed Pervasive Systems. However, the Distributed Computing newlineSystems exist in multiple forms such as Cluster Computing and Grid Computing newlineSystems. The grid computing form of distributed systems raises several challenges as they newlinecover increasingly large networks. Moreover, the assemblage of distributed, newlineheterogeneous resources form what is called as Computational Grids and emerges as a newlinechallenging field in the area of distributed system. In addition, as the resources are newlineheterogeneous in nature, they eventually become slow and unreliable. This addresses the newlineproblems of resource sharing, scalability, etc. Hence, studying the performance measures newlineof such a huge network becomes utmost important. newlineBesides this, after a thorough review of the existing literature related to grid newlinecomputing and allied research, it is observed that the existing approaches suffer from newlineseveral shortcomings such as the reliability of the network which gets accentuated when newlinenumber of nodes increases, the resource allocation of both software and hardware, newlinescheduling and synchronizing to comply within a stipulated time and budget and the newlinesustenance of bandwidth, etc. The research made in this thesis aimed to study and newlineidentify the gray areas of Computational Grid networks that can address the above newlinevii newlinementioned problems. Hence, the objective is to analyse the system network through newlinedifferent optimization techniques and to design its application in Computational Grid to newlineenhance the efficiency of accomplis