Cost Optimization of Virtual Machine Provisioning in Federated IaaS Clouds

Abstract

In this paper, we present cost optimization model in cloud computing, and formulate the cost-aware resource allocation problem that provides cost-efficiency in the context of the cloud federation. Our model assumes a cloud provider with multiple heterogeneous resources or data centers. The provider needs to control amount of resources to avoid overprovisioning and increasing capital costs. To reduce an importance of known Build-To-Peak approach that means building infrastructures for top demands with over-provisioning in total operating time, cloud provider has to collaborate with other providers to be able to fulfil requests during peak demands by using idle resources of other peers. In this scenario, it is important to find a trade-off that allows reducing the total investment and operational cost. We address cost minimization problem in the hierarchical federated cloud environment, where external clouds are parameterized by renting costs per time unit. We discuss several cost optimization algorithms in distributed computer environments with the goal to understand the main characteristic of the cost optimization. We conclude by showing how none of these works directly addresses the problem space of the considered problem, but do provide a valuable basis for our work.