Neural load prediction technique for power optimization in cloud management system

Cloud computing is the current technology used for sharing and accessing resources via internet. It provides a scalable and cost effective environment. Large number of servers in the data centers leads to huge consumption of power in the cloud computing scenario. Optimization of power consumption is a key challenge for effectively operating a datacenter. Power consumption can be regulated by using a proper load balancing technique. Load balancing is done so as to distribute the load fairly amidst the servers and also a scheduling technique k followed to selectively hibernate the servers to optimize the energy consumption. The load balancing is based on load prediction and server selection policy. Neural Network is used for load prediction, which predicts the future load based on the past historical data. The servers can be monitored and given ranking based on their reliability record and this information is used as a criterion while performing load balancing.