PrioritizationofOverflowTaskstoImprovePerformanceofMobile.docVIP

Prioritization of Overflow Tasks to Improve Performance of Mobile Cloud Abstract Mobile devices may offload their applications to avirtual machine running on a cloud host. This application mayfork new tasks which require virtual machines of their own onthe same physical machine. Achieving satisfactory performancelevel in such a scenario requires flexible resource allocationmechanisms in the cloud data center. In this paper we presenttwo such mechanisms which use prioritization:one in which forked tasks are given full priority over newlyarrived tasks, and another in which a threshold is establishedto control the priority so that full priority is given to the forkedtasks if their number exceeds a predefined threshold.We analyze the performance of both mechanisms using a Markovianmultiserver queueing system with two priority levels tomodel the resource allocation process, and a multi-dimensionalMarkov system based on a Birth-Death queueing system withfinite population, to model virtual machine provisioning. Ourperformance results indicate that the threshold-based priorityscheme not only performs better, but can also be tuned to achievethe desired performance level. Architecture: SYSTEM ANALYSIS EXISTING SYSTEM Existing work on resource allocation in mobile cloud computing, but also some relevant research results pertaining tostationary clouds. The proposed resource allocation module describes the virtual machine provisioning module and integration of the two modules. Existing works that either reply on a linear programming formulation or on intuitively derived heuristics that offer no theoretical performance guarantees. PROPOSED SYSTEM Proposed solutions to address the issues of computational power. Personal use is permitted, but republication/redistribution requires mobile devices by offloading computing tasks has been an approach for extending the concept of VM-based clone cloud offloading from LAN surrogates to cloud servers. Proposed a polyn