Optimal cloud resource provisioning: A Two-Criteria Formulation

Abstract

Federated resource provisioning in an on-demand, instantly procurable fashion with the flexibility of a pay as you go model for pricing has led the path for cloud computing to be the computing technology of the future. However, resource provisioning technology needs to be well-supported by appropriate optimization strategies for sustainability purposes. In this paper, therefore, an efficient resource provisioning strategy has been proposed that arrives at optimal provisioning solution with minimal cost and SLA violation rate. To this end, an optimal cloud resource provisioning model has been formulated using the Stochastic Integer Programming (SIP) problem which has been solved by assuming customers' cloud demand for resources as Poisson’s distribution to accommodate for uncertainties pertaining to user demand.