Optimized Resource Sharing for Federated Cloud Services with Desired Performance and Limited OpEx: IEEE Global Communications Conference (GLOBECOM)

The provision of cloud resources to meet user demands in 5G wireless networks is a challenging task due to the high workload predicted to be experienced by cloud service providers (CSPs). Cloud federation has emerged as a paradigm to support CSPs with resource limitations by borrowing surplus resources of other CSPs in periods of high demands. The major concern of each CSP with resource limitations is to borrow resources from other federation participants in such a way that cloud services are provided to the end-users with a desired grade of service (GoS) as well as the overall profit is maximized. This paper proposes an efficient mechanism based on the merchant mode to dynamically facilitate optimal allocation of cloud resources, maximizing the profit of CSPs as well as improving the GoS. The robustness of the proposed optimal scheme is evaluated by comparing it with the heuristic algorithm. The simulation results demonstrate that at each trading window, the proposed optimal scheme outperforms its heuristic counterpart. Moreover, after 50 trading windows, the proposed approach results in 43.5% gain in net profit to CSPs as well as facilitating 3.35% of additional resource.