Game theoretic approaches for revenue sharing in federated cloud

Cloud federation is a new paradigm in Cloud Computing. It helps the Cloud Service Providers (CSPs) to overcome the limitation of their own resources to serve their clients. The CSPs can negotiate the possibility to borrow external resources from each other when their own facilities are not able to satisfy their client requests. Reciprocally, the CSPs may sell some of their unused resources to another CSPs in case of under load. The aim of this paper is to propose and analyse multiple revenue sharing models in a federated environment in the context of the Easi-Clouds, a European ITEA 2 research project. We discuss the specifications and models of Cloud Federation. We present the problem of revenue sharing in the federation and propose a set of properties the solution should fulfil. We propose and evaluate our revenue sharing models suited to the Federated environment with numerical analysis via simulation. We compare the proposed revenue sharing mechanisms and classify them based on different scenarios. Simple revenue sharing models can be adapted in specific cases. However, Game theoretical approaches can be more suitable in some scenarios, with the drawback of being more complex to implement.