OCPDA: A novel approach towards detection of overloaded cloud providers in a federated environment

Abstract

A federated cloud environment helps different cloud service providers (CSP) to improve their performances in terms of Quality of Service $(\mathrm {Q}\mathrm {o}\mathrm {S})$. For a federation to deliver its best, the workloads on its member CSPs should be balanced. Overloading a CSP leads to degradation in performance which is absolutely undesirable. To balance the workloads among the member CSPs of a federation, detection of the overloaded CSPs is important. The proposed model, Overloaded Cloud Provider Detection Algorithm(OCPDA), uses Multiple Linear Regression (MLR) based technique to estimate the loads of CSPs. Extensive simulation has been conducted which shows that OCPDA successfully detects overloaded CSPs in a federation.