Reward-based Markov chain analysis adaptive global resource management for inter-cloud computing

The cloud IaaS provider supports diverse services for users to access big data of the real-time entertainment or the non-real-time working traffic. The IaaS provider builds data centers that include different types cloud resources/equipment, e.g., physical machines, virtual machines, networking, storages, power equipment, etc., and significantly increases cloud cost. An efficient cloud resource management is required for the cloud provider to maximize system reward while satisfying the QoS of various SLAs. This paper proposes a Reward-based adaptive global Cloud Resource Management (RCRM) that consists of three main contributions: the Large-scale and Small-scale traffic Predictions (LSP), Adaptive Cloud resource Allocation, and Maximum Net Profit. The M/M/m/m Markov chain model analyzes the service blocking and the required number of VMs for each request. For maximizing the system net profit, the cloud providers always oversell cloud resources. However, the cost of deploying data centers at different areas in the world is different. This paper adopts the VM migration-in/migration-out and task redirection to adaptively allocate cloud resources among global data centers. Numerical results demonstrate RCRM outperforms the others in dropping probability, SLA violation, violation penalty and net profit. Furthermore, the dropping probability of analysis is very close to that of simulation and justifies the correctness of the proposed Markov chain model.