Green Coding and Energy Efficiency in REST APIs: Empirical Evaluation

Abstract

As there is a growing demand for energy in IT systems, energy efficiency in IT and green software are gaining interest and are being research in growing amounts. Green coding and energy-efficient practices in software are not a brand new topic, but lack specificity regarding REST APIs. This thesis explores the applicability of existing methods of building energy-efficient software in the context of REST APIs through a literature review and an empirical analysis. A literature review was conducted on the current practices of developing green software, identifying methods such as overhead minimization, performance optimization and optimal design choices. A selection of these practices were selected and REST API implementations were built to test the effects of the practices. The effects were empirically tested in the form of power measurements on the reference implementations of the REST API. All of the empirically validated green coding practices exhibited at least some level of energy efficiency improvements. The underlying technology implementation had the greatest effect, whereas employing caching had the least significant improvement in energy efficiency. This work contributes both empirical data and practical insights toward the development and design of more energy-efficient and green REST APIs. Additionally, the thesis identifies points of focus for future research.