A price-performance analysis of EC2, Google Compute and Rackspace cloud providers for scientific computing

Abstract

One of the recently emerging areas in cloud computing is deployment of virtual machines across multiple clouds based on providers' ranking. This involves benchmarking of different cloud providers, development of different techniques for selection of candidate providers and frameworks for ranking cloud providers. Existing benchmarking studies are mostly focused on selection of best-fit cloud provider among a set of cloud providers for a particular set of quality attributes based on industry best standard tools and techniques. However, most of the researches are focused on performance of IaaS cloud providers and price-performance analysis is normally ignored while benchmarking IaaS metrics. In this work, we propose a novel QoS based ranking methodology along with price-performance analysis that can be used as an input for selecting candidate cloud providers. Our proposed mechanism allows cloud consumers to find the most cost effective virtual machines for a given set of user preferences. As a case study, we present performance evaluation and benchmarking results of three major cloud providers: Google, Amazon and Rackspace.