Master Thesis EVALUATING THE ENERGY EFFICIENCY OF A SOFTWARE SYSTEM A PRACTICAL MODEL

Page created by Michiel Cuijpers on 24 Sep 2013 14:58


by Georgios Kalaitzoglou, University of Amsterdam, Master Software Engineering

download here

Abstract:
The information era has been marked by a remarkable increase in resource demand that has been facilitated by the Information and Communications Technology (ICT) in an energy-agnostic way. As a result, Energy and Power have started to evolve into major design and operational considerations.
Despite, the rising concern regarding the ramifications that ICT’s rising energy con- sumption creates, there is still no practical and widely applicable method that is able to determine the energy efficiency levels of a software system. In an effort to fill this gap and give an answer to the research question: "How to develop a practical model that will have the ability to evaluate the energy efficiency of software systems?" this thesis proposes a model to evaluate the different aspects of energy efficiency that are relevant to the majority of modern software systems.
We first propose a conceptual division of energy efficiency into sub-characteristics that are taken under account by this research. We then provide rationale and raise rel- evant questions in order to identify metrics that can be used to quantify those sub- characteristics. Argumentation regarding the metric choice as well as advice on how to act upon their values is provided. Subsequently, an evaluation is performed in order to determine if the model holds any merit, by applying it to the systems of the Software Improvement Group. The results demonstrate that this modeling approach is promis- ing, although more evaluations are needed to come to a tentative answers regarding its value.
The model is intended to be used by stakeholders and SIG consultants in order to eval- uate the energy efficiency levels that a software system exhibits. It gives an understand- ing of how the energy consumption of the system is attributed to its main components, based on the functionality that they deliver, and provides an indication of system spots where energy waste occurs. This enables potential evaluators to discover optimization opportunities and prioritize them more effectively.

editor: Georgios scored a very nice 8,5 for his thesis.


Add a New Comment
Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-ShareAlike 3.0 License