Towards Semantic-Based Cloud Application Management

Abstract

Cloud computing promises to transform applications and services on the web into elastic and fault-tolerant software. To aid at this target, various research prototypes and products have been already proposed. However, especially with respect to the design phase of cloud-based applications, such prototypes do not enable the appropriate composition of cloud services at different levels to realise not only the functionality but also the underlying infrastructure support for such applications. Moreover, most existing prototypes and products lack the appropriate semantics to guarantee that the respective design product is the most suitable and accurate one according to the various types of user requirements posed. To this end, this article proposes a semantic cloud application management framework that addresses the aforementioned issues by relying on ontologies to semantically describe cloud service capabilities and application requirements, on semantic cloud service matchmakers considering both functional and non-functional aspects as well as on a novel cloud service composition approach which is able to perform concurrently service concretisation and deployment plan reasoning, thus catering for the different levels involved in a cloud environment and their respective dependencies by also satisfying all types of user requirements posed. The service composition approach is experimentally evaluated deriving quite promising results indicating that the state-of-the-art is advanced.