Business value-aware task scheduling for hybrid IaaS cloud

Abstract

Since an Infrastructure as a Service (IaaS) provider has limited resources, it faces the challenge of scheduling tasks to meet the peak demand. For this issue, the task scheduling problem for hybrid cloud from an IaaS provider's perspective should be properly addressed. However, existing research does not consider either process tasks with slack temporal constraints or the business value from an IaaS provider's perspective, or it has to specify the type of Virtual Machine (VM) for each task prior. In this paper, we first formalize the problem in which both process tasks with slack temporal constraints and VM type selection optimization are considered, and business value is taken as the objective to comprehensively measure the earned value of an IaaS provider. Second, a novel Evolutionary Algorithm for Task Scheduling in a Hybrid IaaS Cloud (EATSHC) is proposed to solve the problem. Finally, several experiments show our approach performs more effectively and efficiently than existing research.