Mutomvo: Mutation testing framework for simulated cloud and HPC environments

Abstract

Many current applications provide high performance to process large volumes of data. These applications usually run in highly distributed environments, like cloud and HPC systems. Nevertheless, the large and complex architectures required for deploying these applications may not be available during the development phase. This limitation can be overcome by using simulation platforms to model a wide range of distributed system configurations and execute these applications in the modeled system. Usually, these applications are tested against a small number of test cases that are manually designed by the testers. It is desirable to have effective test suites in order to detect failures in the application models. In this paper we propose a mutation testing framework for detecting errors in distributed applications executed in simulated environments. The execution of a test suite against the set of mutated models allows to determine its effectiveness for detecting different errors. The proposal has been implemented in a tool called MuTomVo. In order to support the feasibility of the proposal, we have carried out a case study over three applications running in different distributed systems: a client/server model, intensive computation and scientific pipeline.