On the Detection of Passive Malicious Providers in Cloud Federations

Abstract

Cloud federation has emerged as a new business architecture which aims to help cloud providers cope with the increased waves of demands on their resources and services. Although plenty of solutions have been proposed trying to ensure the optimal formation of cloud federations, these approaches ignore the problem of encountering malicious providers that join federations to destroy them from inside and exclude some strong competitors from the market. To tackle this challenge, we propose, in this letter, a maximin game theoretical model which assists the broker, responsible for creating and managing federations, with maximizing the detection of such malicious providers. The challenge here is to deal with providers that try to minimize the detection maximization through distributing their misbehavior over several federations and changing their identities from time to time. Experiments conducted using real data from the CloudHarmony dataset reveal that our solution maximizes the detection of malicious providers and improves the profit and quality of service of the federations compared with the sky federation model