QoS Instrumentation to Support Cloud Computing SLA Assurance
Résumé
Today, organizations and individuals are moving from proprietary servers to the cloud computing to benefit from the powerful advantages of the theoretical unlimited resources and computation power. The principle of PAYG (Pay As You Go) makes this new paradigm more attractive and provides a rapid growing and extension. However, using a multi-tenancy environment brings new challenges in terms of security, reliability and QoS (Quality of Service). Thus, to fulfill cloud customer expectations in cloudified applications, a relationship must be formalized in a signed contract (i.e. SLA or Service Level Agreement). The service provider monitors the application components to prevent breaches during the SLA life cycle and guarantees the enforcement of the promised QoS. Usually a service is monitored as a whole instead of considering each component in the service individually for a best intervention and an optimized scaling. In this paper, we propose a component-based application monitoring mechanism. We have adopted the OCCI (Open Cloud Computing Interface) framework to model our agreement and instrument a response-time metric for SLA enforcement. The goal is to prevent latency and SLA violation by identifying responsible components involved in the request chain. Our priority is to take actions to prevent the breaches or at least act for a resolution when a problem manifests.
Origine : Accord explicite pour ce dépôt
Loading...