Agent based network monitoring in grid environment

Abstract

The Grid is a heterogeneous collection of resources providing a reliable admittance to all the users thus enabling scalable virtual organizations for resource sharing among the geographically distributed communities. Grid monitoring involves the monitoring of the available computing resources and network characteristics. In this thesis, a four layered Grid network monitoring architecture is modeled with the Grid scheduler in the collective layer. The proposed Grid network monitoring retrieves network metrics using sensors as network monitoring tools. This thesis presents the computing resource monitoring in Grid using available free RAM to select the suitable resource for job submission. This thesis also presents the network aware resource selection strategy by analyzing the network metrics along with resource metrics for effective utilization of resources. The proposed monitoring system is integrated with CARE (Centre for Advanced Computing Research and Education) Resource Broker and tested. The experimental results are evident for the minimization of job completion time for the submitted job when compared with the conventional method. The simulation results also prove that the more number of jobs are completed with the proposed network aware resource selection strategy which improves the better utilization of the Grid resources with high success rate of the submitted jobs. The job status and the progress are updated at the Resource Broker periodically. This thesis also presents History Based (HB) prediction approach to predict the performance of network using network statistics of network metrics. The experimental results exhibit that the proposed prediction measurements closely matches with the measured one. The proposed Grid network monitoring system can be extended with additional network characteristics like router traffic and network availability. The prediction module can also be enhanced further to tune the network parameters for the effective utilization of computing and network resources. newline newline