Modeling energy-aware web services and application. (May 2016)

• Record Type: Journal Article
• Authors: Bartalos, Peter; Wei, Yi; Blake, M. Brian; Damgacioglu, Haluk; Saleh, Iman; Celik, Nurcin
• Is Part Of: Journal of network and computer applications. Volume 67(2016)
• Page Start: 86
• Page End: 98
• Publication Date: 2016-05
• Subjects: Energy-aware computing -- Green web services -- Energy-efficient system architecture -- Architectural support for energy aware computing; Microcomputers -- Periodicals; Computer networks -- Periodicals; Application software -- Periodicals; Micro-ordinateurs -- Périodiques; Réseaux d'ordinateurs -- Périodiques; Logiciels d'application -- Périodiques; Application software; Computer networks; Microcomputers; Periodicals; 004.05; 004
• DOI: 10.1016/j.jnca.2016.01.017
• Languages: English
• ISSNs: 1084-8045

Abstract

Abstract: Cloud computing provides an infrastructure where large organizations can redundantly distribute their software assets across geographically distributed data centers. Consequently, on the account of the importance of energy efficiency, these organizations have to strategically configure their overall operations to leverage power availability as it changes, reduce their overall energy costs, and promote energy-efficiency throughout the entire system. Traditional state-of-the-art approaches rely on the monitoring and configuration of systems based on load measurements taken directly from their hardware assets. These measurements are largely independent of the underlying software applications. In this work, we introduce a decision support procedure to provide a priori, deterministic understanding of power consumption of modular software assets or services that reside on the hardware devices/servers. The proposed decision support procedure relies on power estimation models that predict power consumption of a software service considering the type of server on which it resides. The proposed procedure also embeds a smart sampling technique that will help monitor and record the system behavior effectively for the provisioning of new software services. Experiments demonstrate favorable predictions of the power consumption of a specific web service or group of web services and the promise of more energy-efficient operations of distributed cloud environments.