PRESS: PRedictive Elastic ReSource Scaling for Cloud Systems

Cloud systems require elastic resource allocation to minimize resource provisioning costs while meeting Service Level Objectives (SLOs). In this paper, the authors present a novel PRedictive Elastic reSource Scaling (PRESS) scheme for cloud systems. PRESS unobtrusively extracts fine-grained dynamic patterns in application resource demands and adjusts their resource allocations automatically. The approach leverages light-weight signal processing and statistical learning algorithms to achieve online predictions of dynamic application resource requirements. They have implemented the PRESS system on Xen and tested it using RUBiS and an application load trace from Google. The experiments show that the light-weight resource demand prediction schemes can achieve better resource prediction accuracy with both lower over-estimation and under-estimation errors than previous approaches.