Title: MELA: elasticity analytics for cloud services
Authors: Daniel Moldovan; Georgiana Copil; Hong-Linh Truong; Schahram Dustdar
Addresses: Distributed Systems Group, Vienna University of Technology, Argentinierstrasse 8/184-1, A-1040 Vienna, Austria ' Distributed Systems Group, Vienna University of Technology, Argentinierstrasse 8/184-1, A-1040 Vienna, Austria ' Distributed Systems Group, Vienna University of Technology, Argentinierstrasse 8/184-1, A-1040 Vienna, Austria ' Distributed Systems Group, Vienna University of Technology, Argentinierstrasse 8/184-1, A-1040 Vienna, Austria
Abstract: While cloud computing has enabled applications to be designed as elastic cloud services, there is a lack of tools and techniques for monitoring and analysing their elasticity at multiple levels, from the service level to the underlying virtual infrastructure. In this paper, we focus on monitoring and evaluating elasticity of cloud services, crucial for supporting users and automatic elasticity controllers, to understand the services' behaviour, and to develop smarter mechanisms for controlling their elasticity. We define novel concepts, namely elasticity space for describing the elastic behaviour of cloud services, and elasticity pathway for characterising the service's evolution through the elasticity space. We introduce techniques for enriching monitoring information and determining the elasticity space and pathway. Based on the above, we introduce MELA, an elasticity analytics as a service, providing features for monitoring and analysing the elasticity of cloud services in multi-cloud environments. To illustrate our approach, we conduct several experiments on an elastic data-as-a-service for a cloud-based machine-to-machine (M2M) platform.
Keywords: elastic computing; cloud services; elasticity analytics; monitoring; cloud computing.
DOI: 10.1504/IJBDI.2015.067569
International Journal of Big Data Intelligence, 2015 Vol.2 No.1, pp.45 - 62
Received: 30 Nov 2013
Accepted: 14 Sep 2014
Published online: 21 Mar 2015 *