Cloud computing allows commercial customers Go up and down the use of resources according to your needs. Many of the gains promoted in the cloud model come from multiplexing resource through virtualization technology. We present a System that uses virtualization technology for dynamically allocate data center resources based on applications require and support green computing optimization of the number of servers in use. We present the “Time Slot Filtering” concept to measure the in equality in the use of multidimensional resources of a server. From minimizing the time, we can combine different types of workloads very well and improve the overall use of server resources. We develop a series of heuristics that to avoid effectively overload the system by saving energy used. Trial controlled experimentation and simulation results show that our algorithm achieves a good performance.

Keywords : Data center in the cloud, Greedy heuristic, Automatic learning, Hyper-realistic, Linear integer programming, trade-off of energy network traffic, QoS, Evaluation of scalability, Time Slot Filtering.