Description

It is easier to predict workload when task is not complex but it is difficult to predict grid performance if a task is complex because heterogeneous resource nodes are involved in a distributed environment. Long execution workload on a grid is even harder to predict due to heavy load fluctuations. In this paper we use, polynomial fitting method for cpu workload prediction. While predicting the workload of the grid error may occur during the prediction ,such error are denoted as prediction errors mean square error analysis is calculated, error means difference between the true value and predicted value.The errors are minimized by the technique called EBAF(estimation based adaptive filter method) Finally computing performance is evaluated by benchmark technique . Grid computing [1], the internet-based infrastructure that aggregates geographically distributed and heterogeneous resources to solve largescale problems, is becoming increasingly popular because it provides us with the ability to dynamically link resources together as an ensemble, to support the execution of large-scale, resource-intensive, and distributed applications. Task scheduling of the applications is an important component for achieving high performance in a grid computing environment, while the prediction of execution time of every application is one of the most important elements in determining such scheduling. AGGREGATED grid performance is directly related to the collective workload to be executed on a large number of processors scattered on all participating grid sites.Predicting the collective grid workload is a very challenging task because heterogeneous resources are widely distributed under the control of different administrations. We propose a new adaptive approach to predicting workload on computational grids within a confidence window, which is dynamically trained with the load variations. The grid workload is represented by a collective load index among all processors. The load index XðtÞ is the percentage of processors utilized within a unit time interval ½t _ 1; t_. All discrete-time instances t are denoted by nonnegative integers. For simplicity, we assume 5 minutes per time step. Load index reflects the CPU utilization rate among all available processors in a grid platform. For example, XðtÞ ¼ 0:45 implies that 45 percent processors are busy during the observation period. Workload is difficult to predict due to the lack of runtime information on job scheduling and resource allocation on remote machines [22], [29]. Predicted workload may contain errors, if loading noises cannot be filtered out. Some previous workload prediction methods have ignored two problems: One is the workload measurement errors and another is the load data noise introduced by workload fluctuation