Novel approaches for load balancing in grid using quality of experience and context aware load prediction

Abstract

Grid is a kind of distributed system defined as pool of resources collected from idle systems connected over the network across the world and provide better sharing for the consumers on the request under pay for use It is a system meant for greater availability from the distributed repository which provides a new powerful and innovative platform that caters the need of massively computational or data intensive applications in terms of resources like processor memory input and output devices data storage clusters servers services networking etc It differs from traditional computing systems because of its nature of heterogeneity and background and parallel workloads Utilization and performance of the grid depend on the optimal balancing of load among the available nodes It is very complex and highly dynamic in nature Load balancing is the process of distributing the load to different systems depends upon estimated time fairness current and future loads utilization capacity availability of resources communication overhead environment etc In line with the same computational grid differs from traditional high performance distributed computing systems in the heterogeneity of the computing nodes and communication links as well as background workloads in the computing nodes Finding an optimal solution in load balancing for such an environment using the traditional method is a NP Hard problem whereas heuristic approaches will provide near optimal solutions Algorithms that could capture the dynamic need and complexity have to be developed for solving wide range of load balancing scenarios newline