Vahid Chinaveh - Department of Computer,Sepidan Branch,Islamic Azad university,Sepidan,Iran
Amin Tousi - Department of computer,Shiraz Branch,Islamic Azad University,Shiraz,Iran

By daily and excessively expanding the number of users connected to the global network,using the online and network-based services is increased. By expanding the number of users whoare online and connected to the internet, cloud computing has made this capability simpler byproviding different services. Therefore, the cloud computing has been accepted, in one hand, bythe investors and large information technology companies and, in another hand, by the internetusers. Hence, the number of the cloud data centers which are providing different services to usersover the world is annually increased. These data centers are made up of hundreds ofcomputational hosts and tens of switches and other network equipment and they daily consumemore energy. However, the investors of the cloud computing try to reduce the costs (includingthe power consumption) to have more customers and they finally get more profits. Differentmethods like the static and dynamic management of the virtual machines and the automaticprocessor frequency adjustment system have been provided for reducing the energy. In thisthesis, a risk-based dynamic cloud resource management algorithm is proposed for decreasingthe energy of the data centers and increasing the profits of the providers and the results areimplemented and evaluated by the CloudSim simulator. The obtained results demonstrate asignificant decrease in the energy consumed by the hosts and eventually, an increase in the profitby considering the amount of the service level agreement violation.

Cloud, Data Center, Virtual Machine, Service Level Agreement Violation, Service Quality