DEEPC: Dynamic Energy Profiling of Components

Liu, Xiaodong; Chinenyeze, Samuel J.; Al-Dubai, Ahmed

Abstract

Many software projects are built using reusable components (i.e. reusable objects - as per component and connectors in software architectures). During component selection in CBSD, components are evaluated on the criteria of required quality attribute prior to integration into a system. Current green software research exploring software energy efficiency as a quality attribute adopts conventional counter-based white box energy measuring approach. Although the conventional approach provides results at fine granularity, as with its adoption in component selection, the challenge is that to test software energy of each component, the test has to be done prior to integration, which means implementing multiple counters or multiple versions of the system -- thus inefficient, especially when involving much components. In this paper, we present an approach and tool for dynamic energy profiling of components for software systems (DEEPC). The proposed approach employs AOP concepts to expedite the energy measurement of components and improve accuracy, by, i) dynamically loading related components for evaluation (load-time weaving) into the base system, as a way to circumvent manual counter implementation, ii) using pointcuts to facilitate power measurement of loaded components. An evaluation of DEEPC approach presents it to be more time and resource efficient with better profiling accuracy compared to its counterpart.

• Presentation Conference Type: Conference Paper (Published)
• Conference Name: 2016 IEEE 40th Annual Computer Software and Applications Conference (COMPSAC)
• Start Date: Jun 10, 2016
• End Date: Jun 14, 2016
• Acceptance Date: Apr 30, 2016
• Online Publication Date: Aug 25, 2016
• Publication Date: Jul 31, 2016
• Deposit Date: Jan 30, 2017
• Publicly Available Date: Nov 15, 2017
• Publisher: Institute of Electrical and Electronics Engineers
• Series ISSN: 0730-3157
• Book Title: 2016 IEEE 40th Annual Computer Software and Applications Conference (COMPSAC)
• ISBN: 9781467388450
• DOI: https://doi.org/10.1109/COMPSAC.2016.90
• Keywords: Green Software, Component Energy Evaluation, Software Energy Profiling, Dynamic Energy Evaluation, Aspect-Oriented Profiling
• Public URL: http://researchrepository.napier.ac.uk/Output/451247
• Contract Date: Nov 15, 2017