Optimizing Quality of Resilience in the Cloud
Authors
Abstract
Quality of Resilience (QoR) is natively connected to cloud computing to enable high-availability of services and secure access to information. QoR, in such dynamic contexts, cannot simply rely on reactive protection models, such as the primary- backup model supported by the Stream Control Transport Protocol (SCTP). Indeed, proactive mechanisms are required to enable an optimal QoR.In this paper we propose the use of multiple criteria that assess the performance of services operating in the cloud to trigger proactive resilient mechanisms. Such criteria, considering the maximization of benefits and decrease of costs, are combined through Multiple Attribute Decision Mechanisms (MADM) to optimize quality of resilience, going beyond mechanisms relying simply on network criteria. In this context, TOPSIS and MeTH are implemented to enhance data transfer services.
Evaluation results, from the tests performed in the cloud environment of a Telecom Operator, demonstrate that MeTH is able to enhance high volume data transfers in comparison to TOPSIS and also to improve resilience support of SCTP, in the presence of different kinds of failures, network conditions and server loads.