Journals Proceedings

Abstract

Grid Computing is a form of distributed Computing that has emerged as a viable solution to meet the ever increasing needs for computational power and data management capability. Designing solutions in such grid computing framework entails addressing much more complicated issues compared to chore software development, namely concurrency, heterogeneity, scalability and so forth; just to name a few. In order to simplify the task of programming in grid environment a software layer is employed to mask off the massive underlying heterogeneity in network, hardware, operating system and programming languages, known as the middleware. Moreover, resources in a grid are dynamic and thus incorporating appropriate scheduling mechanism becomes a challenging proposition. This paper addresses some major issues in context of job scheduling in computational grids: namely, average active jobs, busy time of CPU, heap memory and average CPU load. They are treated as Work Load Description (WLD) in a grid scenario. For experimentation purposes, Grid Gain has been incorporated as middleware. Experiments have been conducted and subsequent results presented herein demonstrate the efficacy of Grid Gain as a platform of implementation of grid computing for catering to future generation computational needs, including load balancing.