A Hybrid Approach for Process Scheduling in Cloud Environment Using Particle Swarm Optimization Technique

Abstract

Resource Management is a central aspect in Cloud Computing. This is because the Cloud Services are offered to several users, in a way which ensures rapid deployment and on demand availability. The primary motivation behind cloud computing is to access desired resource over cloud/internet, using thin-clients capable of running only an internet browser program. A given physical resources is extended as multiple virtual resources which can be allocated to several users through virtualization paradigm. One a particular requirement is completed; the corresponding virtual resource can be allocated to another user or shall be stopped for execution to save energy/ enhance throughput. Resource management in a scenario, consisting of large population of users, over a number of resources, with various priorities, under different load conditions is a complex optimization problem. This issue has been discussed at length in research literature over different service models like SaaS, PaaS and IaaS. In this paper, a generic model of cloud service facility is considered which falls under the category of public cloud. Nature inspired algorithm, particularly, particle swarm optimization is presented to manage the request-response architecture, for minimum latency. The proposed approach outperforms the benchmark approaches presented so far. The simulation is performed on CloudSim and the results obtained are in excellent agreement to the analytical model.