Cloud computing architectures and dynamic provisioning mechanisms

Cloud computing is a prodigious modern technology committed to provide pool of resources to the on-demand customers. The resources in a cloud are virtualized as a collection of services using the virtualization technology. Efficient provisioning of resources is a lofty problem due to the lively nature and the need to support heterogeneous applications with diverse performance requirements. The performance guarantee from the cloud datacenter requires efficient utilization of the resources. The efficient resource utilization towards any specific Service Level Agreement (SLA) or constraint on Quality of Service (QoS) alone is not sufficient for the cloud computing environments. Thus, cloud computing requires to strike a balance between the performance based on the negotiated QoS as well as the energy consumption at the datacenter during the process of resource provisioning. In this paper, we provide a detailed review of cloud computing architectures and provisioning mechanisms for delivering the computing as a service. We classify the cloud computing architectures discussed in literature based on their significance in cloud services provisioning mechanism. The paper presents taxonomy of dynamic provisioning mechanisms from the cloud utility point of view and brings out salient features/evaluation of existing mechanisms.