Cost-Optimized Resource Provision for Cloud Applications

With an increasing number of cloud providers starting to offer virtual resource leasing services, application providers have more choices when requiring resources. But how to achieve the most cost-optimized resource solution is still a challenge. To address this problem, we propose a cost optimized resource provision approach which achieves the cost saving target for application providers. In our approach, we figure out the projection between workload and VM amount according to predicted workload. By exploiting pricing policies of existing cloud providers, our approach presents a hybrid resource provision solution, which involve instances of different prices and leasing durations. The experiment results demonstrate that our approach achieves the cost-saving goal with few SLA violations and little resource waste.