Resource Allocation under Uncertainty in Cloud Storage for Computational Environment
A global computation market could be realized by a high-performance federated architecture that has cloud computing providers. They use economic aware allocation mechanisms driven by the underlying allocation requirements of cloud providers. In cloud, resources are reserved by m participants during auction and in most cases the resources are utilized by n users. Therefore all other m-n reservations are wasted in the duration of auction. In this process last minute bidding is used which is time and cost expensive. To overcome this wastage we propose two principles in general, first avoid commitment of resources, second avoid repeating auction and allocation process. We have distilled these principles into five high-performance resource utilization strategies, namely: overbooking, advanced reservation, just-in-time (JIT) bidding, progressive contracts, and using substitute providers to compensate for encouraging oversubscription. These strategies are examined experimentally with the DRIVE Meta scheduler. Several diverse synthetic workloads have been used to measure both the performance benefits and economic implications of these strategies.