Mechanism design for robust resource management to false report in cloud computing systems

Abstract We consider a resource allocation problem that ensures a fair QoS (Quality of Service) level among selfish clients in a cloud computing system. The clients share multiple resources and process applications concurrently on the cloud computing system. When the available resources are less than the total amount of required resources by all clients, the overload condition occurs. To avoid this, a fair resource allocation is needed. However, when there are selfish clients who want to maximize QoS levels of their applications, they may not report their true QoS functions honestly in order to get more resources than their fairly allocated ones. Then, the performance of the system degrades. Thus, it is important to prevent selfish behaviors of the clients. We propose a resource allocation mechanism that ensures a fair QoS level based on the framework of the mechanism design. In the proposed mechanism, the resource manager cannot know applications which will be processed by the clients but can observe their QoS levels after completing the applications.