Scheduling Work Load Based on Priority in Cloud

Abstract

Cloud computing offers ability to provide parallel and distributed simulated services remotely to the users through the internet. Services hosted within the cloud can potentially incur processing delay due to load sharing among other active services, and can cause active optimistic simulation protocols to perform poorly. Number of complex application runs in remote data centres, parallel processing capabilities often show a increase in utilization of CPU resources as parallelism grows, mainly because of communication and synchronization. To achieve certain level of utilization, Our proposed method partitions a nodes computing capacity into the 4-tiers with low CPU priority, medium CPU priority, high CPU priority and very high CPU priority. In large data-center, processes of a job may need to be allocated to nodes that are close to each other to minimize the communication cost. We provide scheduling algorithms for parallel jobs to make efficient use of the k-tiers VMs to improve the responsiveness of these jobs. We focus on improving resource utilization for data-centers that run parallel jobs, particularly we intend to make use of the remaining computing capacity of data-center nodes that run parallel processes with low resource utilization to improve the performance of parallel job scheduling. The method is practical and effective for consolidating parallel workload in data centres.

Keywords: Distributed Computing, Parallel Computing, Parallel Simulation, Resource Consolidation, Scheduling, Virtualization

View PDF