Cloud Resource Allocation; Auto-Scale Data Centers Issue

Cloud Resource Allocation; Auto-Scale Data Centers Issue

Cloud computing paradigm provides on-demand access to the computing resources based on virtualization technology through the web. Users can lease the resources instead of owning the expensive resources and get rid of the troubles for the maintenance of equipment. This leads to transforming the resources to cloud environments and clarifies the importance of resource provisioning issue for cloud service providers. The problem of resource allocation for requested services reduces to how to place the virtual machines on physical hosts in infrastructure as a service cloud model. Optimal performance of this process must guarantee the quality of service as mentioned in the service level agreement contract. In this paper, we propose an architecture of modeling for the performance of a data center in the cloud environment using the birth-death process of the queueing system. We assume that jobs arrive in the system, according to the Poisson process and their service times are exponentially distributed. The load of the data center can be kept optimal while the number of hosts that are in running state will be minimum in this model. A new host will be added when the utilization of current host reaches the predefined threshold and a host will be released while the utilization of the host decreases so much that it can be eliminated from active hosts list. We have presented numerical results for a sampled system.

Cloud computing paradigm provides on-demand access to the computing resources based on virtualization technology through the web. Users can lease the resources instead of owning the expensive resources and get rid of the troubles for the maintenance of equipment. This leads to transforming the resources to cloud environments and clarifies the importance of resource provisioning issue for cloud service providers. The problem of resource allocation for requested services reduces to how to place the virtual machines on physical hosts in infrastructure as a service cloud model. Optimal performance of this process must guarantee the quality of service as mentioned in the service level agreement contract. In this paper, we propose an architecture of modeling for the performance of a data center in the cloud environment using the birth-death process of the queueing system. We assume that jobs arrive in the system, according to the Poisson process and their service times are exponentially distributed. The load of the data center can be kept optimal while the number of hosts that are in running state will be minimum in this model. A new host will be added when the utilization of current host reaches the predefined threshold and a host will be released while the utilization of the host decreases so much that it can be eliminated from active hosts list. We have presented numerical results for a sampled system.