Grid Computing has become an important field of research, which has evolved from conventional Distributed Computing and High Performance Computing for solving large-scale problems. Scientific and business applications are very complex and it requires massive computing power and storage space. Grid Computing environment supports the technology to execute large-scale applications on Resources. Resource Allocation and Task Scheduling have received much attention from the research community in the Grid computing because of some desirable properties like optimum utilization of resources, improvements in minimize of waiting time, total completion time and total response time and Total Resource Cost. The scheduling of tasks for the heterogeneous computing resources has been examined by many scientists and has been established to be an Non deterministic Polynomial complete problem.
Numerous methods and strategies have been developed by many researchers to perform Resource Allocation to the assigned tasks in Grid Computing Environment. However, major problems still exist in the literary works that are inefficiency in Resource Allocation based on economic scheduling and resource insufficiency for executing jobs in conventional methods which degrade performance of Grid system. Hence it is proposed “Novel Strategies for Resource Allocation and Scheduling in Grid Computing Environment” to increase the performance of a Grid system with an objective of maximizing of utilization of resources and minimizing of waiting time of a job in job pool and makep.
The main contribution of this work is proposing Resource Allocation with penalty function using Genetic Algorithm which consists of two allocation models i.e. Allocation With Out Penalty Function (AWOPF) and Allocation With Penalty Function (AWPF). However, both the models have used penalty function where as one of the allocation models considers economic significance. The proposed models are compared with the conventional resource allocation models like Swift Scheduler (SS) and First Come First Scheduler (FCFS) in terms of utilization of resources, Total resource cost and makep. Both the models have used Genetic Algorithm (GA) to find appropriate resources. The analysis has proved that the proposed models are strong enough even under all conditions.
Genetic Algorithm is a widely used approach by the researchers for solving NP-complete problems. Even though GA is used widely but the main drawback is slow to solve scheduling issues in realistic environment due to its number of iterations. In this context, it is proposed an Optimized Genetic Algorithm for Resource Scheduling in Grid Computing which can reduce the search time by limiting the number of iterations and improves the convergence rate. At the same time feasible solution can be obtained for Resource Scheduling by resembling the genetic process.
In Grid Computing, Tasks are always organized as workflows. Workflows of Scheduling is a main problem in grid computing in terms of certain requirements and also it influences the performance of the grid system. A few algorithms in literature are implemented which deals with scheduling workflows, but most of them are concentrated on single parameter or with small scale workflows but not suitable for large scale workflows. In this point of view, it is proposed a Hybrid Genetic and Ant Colony Optimization (GAACO) algorithm which is a combination of Genetic algorithm and Ant Colony Optimization (ACO) algorithm to solve large scale workflows. This algorithm schedules large scale workflows with different parameters. Experiments are carried out by small, medium, large grids using Grid Simulator and results have proved that the efficiency of the proposed algorithm has been improved.
The grid resources are belonging to different domains and are distributed in different geographical regions, decentralized method is an appropriate solution for Resource Management in Grid Computing Environment. A suitable Resource Management is required which exploits the resources effectively and satisfies the customer requests. To satisfy the customer requirement Genetic-Auction based algorithm (GAAB) has been proposed to allocate resources to the tasks with the idea of Genetic Algorithm and Microeconomics.
This algorithm contains two modules, Auction module and Genetic Algorithm based module. Auction module find outs resource trading price between resource provider and resource buyer. The Resource Allocation carried out by Genetic Algorithm based module by considering both time and cost constraints simultaneously. Evaluations are done using simulation environment and the results indicate the effectiveness of the proposed model.
From the above contributions can conclude that the proposed methods can improve the performance of grid system by maximizing resource utilization when compare to the conventional methods.