A Self-Tuning Procedure for Resource Management in InterCloud Computing

InterCloud Computing is a new cloud paradigm designed to guarantee service quality or performance and availability of on-demand resources. InterCloud enables cloud interoperability by promoting the interworking of cloud systems from different cloud providers using standard interfacing. Resource management in InterCloud, considered as an important functional requirement, has not attracted commensurate research attention. The focus of this paper is to propose a Software Cybernetic approach, in the form of an adaptive control framework, for efficient management of shared resources in peer-to-peer InterCloud computing. This research effort adopts cooperative game theory to model resource management in InterCloud. The space of cooperative arrangements (resource sharing) between the participant cloud systems is presented by using Integer Partitioning to characterise the worst case communication complexity in peer to peer InterCloud. Essentially, this paper presents an Integer partition based anytime algorithm as an optimal cost solution to the bi-objective optimisation problem in resource management, anchored principally on practical trade-off between the desired performance (quality of service) and communication complexity of collaborating resource clouds.