Search

Cloud computing is emerging as a new provision model, enabling ubiquitous, convenient, on-demand network access to a shared pool of configurable computing resources. It delivers a high capacity IT infrastructure according to different models. Notwithstanding cost saving bene ts of ownership of IT investements especially with multi-tenancy settings, the explosion of configurable Cloud services and resources on the Internet and the huge number of Cloud tenants raises new research challenges in multi-tenancy aware configurable Cloud services discovery and selection. In this paper, we propose to employ variability modeling as a tool to consider relevant aspects of Cloud services. In particular, we use extended feature modeling to abstract the commonalities and differences of Cloud services. Based on our modelling, we propose an approach for the discovery and selection of infrastructure Cloud services. Recent research works have been devoted to perform the configuration search process by listing and matching an exponential number of configurations resulted from a single Cloud service causing a large potential search space that lead to scalability issues. In our approach, to overcome these issues, the problem of configurable matching is reduced to a graph matching problem and graph edit distance algorithms are adapted to this purpose. To ensure QoS awareness, we use directed weighted graphs algorithms to rank the best matching instances. Experimental results demonstrate that the proposed algorithms can garantee the accuracy and effectively improve the scalability of Cloud infrastructure service discovery and selection. [ABSTRACT FROM AUTHOR]