Agent-Based Intrusion Detection for Federated Clouds

In the last years, the cloud services market has experienced an extremely rapid growth, as reported in several market research reports, which may lead to severe scalability problems. Therefore, federating multiple clouds is enjoying a lot of attention from the academic and commercial point of views. In this context, publish-subscribe is a widely used paradigm to support the interoperability of federated clouds. In this paper, we describe some potential vulnerabilities of a publish-subscribe based federated cloud system. In particular, we propose an agentbased system that aims at monitoring security vulnerabilities that affect such kind of inter-cloud cooperation solutions.