Efficient Agent Based Priority Scheduling And Loadbalancing Using Fuzzy Logic In Grid Computing
i-manager s Journal on Computer Science · 2015
Grid computing is a kind of distributed computing in which the resources are organically scattered and controlled by the grid manager. It is used to solve the large scale integration problem in scheduling and load balancing. Grid can be classified into two types; they are Data Grid and Computational Grid. In Data Grid, large amounts of organically distributed data individually accessed, modify and transfer for research purpose [1]. It located the data in single or multiple sites for managerial to administrate the data access. The middleware of data grid is used to handle the incorporation among users and the data, also offer to control access to the available data interested in the grid system. Computational grid split the task into multiple parts, and executes them on different grid nodes in parallel, so computation will perform earlier. To improve scalability and fault tolerance of the grid system, the grid user can relieve the computations onto most present nodes. Recent grids must give the list of services based upon the requirement of the individual users. So the grid managers have only limited control over the system components. The resource availability and resource provision is the major problem in the grid environment [2], [17-19]. To overcome this problem, a smart resource and grid tasks managing techniques are essential in grid computing. Grid tasks are managed by using some scheduling methods. The main aim for scheduling is to attain better resource exploitation, secure requirement of the grid users and reduce the energy devoted by all the resources. The grid schedulers should also capture the complexity to the grid system and provide significant measures for a wide range of grid applications and services. Grid scheduler makes the use of different priority and time stamp based scheduling algorithm, to schedule the task. After scheduling the task to corresponding resources, grid scheduler must check the load balancing conditions [3], [20]. Load balancing is a technique, which is used to allocate the workload across multiple grid resources to accomplish the best possible resource utilization and avoid overload. Based on the processing capabilities of the entire node, the workload can be distributed; therefore the time taken to perform all tasks can be minimized. The load balancing technique preserves a static or dynamic nature. In static load balancing algorithms, tasks