A performance evaluation model for users' satisfaction in federated clouds

Cloud federation has been proposed as a new model enabling Cloud providers to cooperate and exchange resources among multiple Clouds to provide uninterrupted services and fulfil the requirements of Cloud users. To sustain the many benefits offered by this environment, it is crucial to have an accurate performance evaluation model. However, evaluating the performance of the federated Clouds environment is challenging for several reasons: the dynamic nature of this environment, the exponential growth of users and the diversity of requests. Therefore, to better satisfy user requests in the federated Clouds environment, we propose in this paper a performance evaluation model using an open Jackson network. Our model considers the key characteristics of the Cloud, such as the diversity of Cloud services and their exponential growth (IaaS, PaaS, and SaaS), as well as the increasing number of user requests, to obtain essential performance metrics, including response time and utilization. These two parameters are obtained analytically and are used to quantify the elasticity of the federated Clouds environment. Several experiments were carried out to evaluate the proposed solution. The results showed that our model outperforms the state-of-the-art and it is usefulness for designing real cloud computing systems.