A Scalable Benchmark as a Service Platform

Abstract

Abstract. Load testing has always been a crucial and expensive activity for software companies. Classical solutions are a real burden to setup statically and their cost are prohibitive in terms of human and hardware resources. Cloud computing brings new opportunities to stress application scalability as load testing solutions can be provided on demand by the cloud. This paper describes a Benchmark-as-a-Service solution that scales automatically the load injection platform and eases its setup according to load profiles. Our approach is based on: (i) the virtualization of the Benchmarking platform to enable the injector's self-scalability, (ii) an online calibration mechanism to characterize injector capacity and impact on the benched application, (iii) a provisioning solution to scale the load injection platform sufficiently ahead of time. We also report experiments on a benchmark that shows the benefits in terms of cost and resources savings.