Amalgamation of Neural Network and Genetic Algorithm for Efficient Workload Prediction in Data Center

摘要

Cloud computing provides every organizations with a plan of action for more effective decision making; expedite revolution, also multitude of other ease. But cloud is further called for attentiveness and problem solving, time observing for management of cost. A present day study on cloud amid IT and Finance executive begin that dearth of tediousness is the sole substantial cloud cost management agony. Respondents mentioned the call for clarified low billing and more precise and predictable budget forecasting. Resource forecasting has become consequential for cost management. We know data centers are spending huge amount for virtual machine migration. The numbers of servers are increasing tremendously. Currently, there are 100 million servers globally where Google and Microsoft own maximum of it . The unnecessary migration can be avoided with workload prediction and turn out the obsolete, idle servers and turn down power consumption. Server consolidation and virtualization are contributing effectively for resource management . The load balancing, reducing the power consumptions by servers, and resource management are few myriad of doing forecasting of resource. Many researchers have proposed various algorithms for prediction, but still reaching 10% accuracy is still challenging. Therefore, it is imperative to accurately predict the workload for resource requirement fulfilling in datacenter. Prediction can be done for long term (1 h) and also for short term (5 min). The accuracy of prediction will be high in long-term prediction.