This In this work, novel approach of using feature selection algorithms for identification of performance features of big data platforms is done. The services provided by big data platforms can be optimized by fine tuning these identified performance features. Service level agreements defined for platform as a service derives its basics from software quality models. This paper focusses on the key role of the performance features in the service level agreements. The research work has identified empirically performance features of Big Data platforms like Hadoop, Spark and Storm. The service level agreements are defined over ISO 25010 quality metrics. The paper reiterates the dependency of service level agreements on software quality model metrics and data platform system parameters for service optimization.
Volume 11 | 11-Special Issue