Asynchronous policy-based grid resource management mechanism for Qos-constrained applications

Abstract

Grid computing [1][2][3] enables the virtualization of distributed computing and data resources such as processes, network bandwidth and storage capacity to create a single system image, granting users and applications seamless access to vast IT capabilities. It brings together all heterogeneous resources and allocates them efficiently to applications. As Grid grows, the complexity of resource management in Grid dramatically increases. To manage the Grid resources efficiently, Policy based resource management system is suitable to the complex management environments. The condition of the grid resources with time space has an uncertainty and a dynamics and it changes. So the policy which made at a moment could not be appropriate to other time. We need some mechanisms to solve this problem. In this thesis, we propose asynchronous policy-based Grid resource management mechanism for maintaining consistency between policy of Grid plane and policy of network plane. Policy decision scheme for management cost and QoS guarantee make available resource quorum with Service Level Agreements (SLA), grid resource and network resource information in advance. However, as the time changes, discordance exist between two policies. It means that the available resource Quorum can not guarantee the QoS service at the other time. If the available resource could not guarantee the user``s SLA in terms of completion time, service management suite divide the job into some sub-jobs to provide QoS service, and then process each sub-job. Through the adjustment scheme for maintaining consistency of policy, we can guarantee QoS service and stable service. In order to evaluate the asynchronous policy-based Grid resource management mechanism we run arterial pulse wave detection simulation which requires QoS service in our testbed. The results show that proposed scheme guarantees QoS for grid applications.