Defining an elasticity metric for cloud computing environments

Abstract

Elasticity is a key property of cloud computing environments and one of the features which distinguishes this paradigm from other ones. An elasticity metric could be used to define and to monitor Service Level Agreements (SLAs), to compare and to benchmark different cloud providers or to improve provisioning and management decisions in real time if it could be measured with some simple procedure, to mention only some examples. Unfortunately, there are not still standard elasticity metrics capable of easily capturing its main components or aspects, these are, scalability, accuracy, time and cost. This work defines a new elasticity metric for cloud computing environments which not only captures these four essential components but also provides a simple procedure to analyse elasticity in cloud contexts. The obtained experimental results on a real cloud environment evidence that this new elasticity metric is able to quantify the degree of elasticity obtained in different cloud scenarios and to establish interesting comparisons and analysis.