Cloud resource provisioning for Amazon EC2

The Cloud Computing can be used as a buzzword for the big turn to an IT world where computing resources provided over the internet can be acquired from IaaS providers with different purchasing plans. A critical challenge is to determine the right amount of resources to minimize the completion time and financial cost from consumers point of view. Finding optimal provisioning solutions thus becomes an NP problem. Taking Amazon EC2 for instance, it provides three purchasing plans with different price and yields different benefits to consumers. This paper addresses the issue of how cloud consumers could make resource provisioning plan for their computation. Heuristic algorithm is devised to determine number of instances required to execute workload so as to achieve optimization goal based on the characteristics of workload and purchasing plans provided by Amazon EC2 which can significantly reduce the total cost incurred to consumer.