查看文章

Cloud Computing has caught the fancy of the research community, big technology companies, application developers and consumers with its promise of on-demand computing and an intuitive service delivery model. While large corporations like Google have invested in creating their million-server warehouses to cater to the tremendous expected demand for cloud resources, smaller service providers may not have enough resources to cater to the “elasticity” inherent in the cloud model. The ability to dynamically provision resources in the face of a volatile resource requests is one of the key performance indicators for cloud service providers and remains a challenge. It is always theoretically possible that the resource requirement exceeds the physically available resources, especially when flash-crowd scenarios are factored in. Catering to peak expected resource requirements by provisioning surplus …