Modeling Performance of Elasticity Rules for Cloud-Based Applications

Many IaaS providers allow cloud consumers to define elasticity (or auto-scaling) rules that carry out provisioning or de-provisioning actions in response to monitored variables crossing user-defined thresholds. Defining elasticity rules, however, remains as a key challenge for cloud consumers as it requires choosing appropriate threshold values to satisfy performance and cost requirements. In this paper we propose novel analytical models that enable the study of application performance under different elasticity rules. Based on these, we develop algorithms for performing scale-in and scale-out operations. We simulate our models and algorithms using different thresholds, and validate the results against empirical data obtained using the same rules with the TPC-W benchmark on Amazon cloud.