Stabilising Performance in Cloud Services Composition Using Portfolio Theory

Abstract

The increasing number of services available in the cloud market make them plausible and attractive for building Cloud Service Compositions (CSC). However, performance instability is common in the cloud environment due to changes in supply and demand of shared computational infrastructure and resources. Candidate compositions are vulnerable to such instability. We propose a novel approach to improve performance stability by leveraging on the principles of design diversity in service composition(s). The approach uses portfolio theory to construct a diversified composition of candidate services that share lowest possible correlation for their performances. We use an exemplar to illustrate the applicability of the approach. Controlled experiments are used to test the approach effectiveness in improving the performance stability of CSC. While the scalability of our approach is evaluated, we report on its sensitivity and effectiveness under multiple correlation settings.