Research On Adaptive Scheduling On Infrastructure As A Service Cloud Computing Platform

IaaS cloud provides a variety of IT capability, such as computing capability, storagecapability, etc., in forms of virtual machines using virtualization of physical resources.Several primary characters of cloud, such as on-demand self-service, resources poolingand rapid elastic, indicates that the quality and quantity of VM services must be adaptedwhen requests arrive. The self-adaptiveness is required by IaaS cloud to ensure theproper actions cloud be enforced automatically to provide a better use experience ofservices.This paper investigates the self-adaptiveness of IaaS cloud, which refers to that theIaaS cloud should keep itself running in the ideal condition by adjusting the structure,modes, deployment of itself without external intervention according to its currentsituation.Based on a variety of assumptions, this article proposed self-adaptiveness schedulingmodels for IaaS Cloud. Specifically, the mathematical expressions and relatedalgorithms of this model are described. The cost of VM migrating is considered, whichmakes the generated scheduling plan practical.According to the experimental result, the proposed model is capable to constructscheduling plan with a relatively accuracy with less system resources consumptions.The consideration of the cost of migrating VM let the generated scheduling plan achievethe goal of a low cost. Meanwhile, the computing complexity of generating schedulingplan via this proposed model is acceptable, which enables the model to be applied in alarge cloud. In sum, the objective of adaptively adaption, including self-adaptivenessof power, performance, resource allocation and structure in IaaS cloud can be achievedby this proposed model.