Data center operators are compelled to deliver an optimal response time to users, as the time taken by the server to respond to changing requirements often impact the Service Level Agreements (SLAs). In several instances, Dynamic Provisioning has proven to be successful in handling gradual changes and maintaining response time. However, time delay is an inevitable factor when operators handle abrupt spikes in data load or unexpected requests for additional capacity. The system warrants time to identify the requirement and set up additional capacity. But, even a delay of few seconds can hamper performance and result in violation of SLAs. At present, data center operators over provision the capacity requirements even when the center is handling normal load, which is an expensive approach. It is essential for the operator to scale capacity based on current requirements for several reasons. Firstly, the increasing emphasis on greener data center creates a need for low-power consumption. Hence, it is essential to bring down power usage by turning off servers that are not in use.
展开▼