An empirical evaluation of database software features on energy consumption

Abstract

Although software does not consume energy by itself, its characteristics determine which hardware resources are made available and how much energy is used. Therefore, energy efficiency of software products has become a popular agenda for both industry and academia in recent years. Designing such software is now a core initiative of software development companies aiming toward social responsibility. Meanwhile, however, developing environmentally sustainable software products is a challenge in that performance, functionality and energy consumption can reflect conflicting goals. In this paper, our objective is to analyze the effects of different features on energy consumption of the IBM DB2, a commonly used database product. The empirical work focuses on three features. We executed a workload in preconfigured software with some features enabled or disabled and with different numbers of users. To compare the different scenarios, three sets of green metrics were utilized. The metric set identified various parts of the software system where energy is consumed. Our findings may suggest that the conflicts among software system performance, functionality, and energy consumption can be mitigated by choosing a combination of features that interact in a way that improves energy efficiency. Index Terms energy consumption, green metrics, energy efficiency, environmental sustainability, software feature interaction, database.