A Proposal of a Simulation-Based Approach for Service Level Agreement in Cloud

Abstract

Clouds offer resources as services, which are charged on the basis of their actual usage. An approach that is currently spreading, to face possible problems linked to a low quality of service offered, is the adoption of Service Level Agreements (SLAs). In this paper we propose the adoption of a simulation-based approach for supporting a framework in charge of SLA management in cloud applications. The idea stems from the consideration that to face the inevitable impact on QoS of the cloud elasticity, it is necessary to evaluate in a predictive way the performance properties of the many different configurations among which the automatic cloud management can choose to allocate resources to customers. This paper proposes the architecture of the framework for SLA management in clouds, identifying the requirements implied by the simulation to be performed and suggesting the adoption of a simulation engine that fits with the proposed requirements.