A FRAMEWORK FOR USER PRIORITY GUIDANCE BASED SCHEDULING FOR LOAD BALANCING IN CLOUD COMPUTING

摘要

Cloud computing has emerged as a new model of computing based on Internet. A cloud with large shared pool ofresources provides computing resources on demand in pay as you go. wo important aspects must be considered: i) from the userperspective latency must be decreased and tasks to be completed as expected, ii) from the service provider perspective, effectiveutilization of computing resources is vital. Many existing solutions that focus on optimized scheduling for improving load balancingdo not consider user priority guidance. When user-guided priority is considered, the scheduling follows well informed approachand it can lead to more efficient load balancing. Therefore it is important to have scheduling algorithms that are guided by userpriority for better optimization of load balancing. In this paper we propose an algorithm known as User Priority based Schedulingfor Load Balancing (UPS-LB) which divides user tasks into elastic and inelastic groups and schedule them for enhanced loadbalancing. In fact, the algorithm makes rescheduling decisions for heavy loaded resources to improve efficiency of load balancingin cloud computing. We built a prototype application to demonstrate proof of the concept. Our empirical study revealed that theproposed UPS-LB algorithm has comparable performance improvement over its predecessors such as Min-Min, LBIMM and PALBIMM.