Characterization of Distributed Cloud Computing Network (DCCN) Server Resource Pool Using Virtualization Dynamics

Distributed Cloud Computing Network is an important subsystem in an on going research referred to as Smart Green Energy Management System (SGEMS). In this paper, we propose a dynamic service availability architecture using two server types (virtualized and legacy server types) and their QoS metrics (particularly service delay time and resource availability) for detailed analysis. The work is aimed at determining the effect of virtualization in the context of resource allocation/scheduling in a cloud based server domain. This was carried out to facilitate the deployment of a previous work on Enterprise Energy Tracking Analytic Cloud Portal (EETACP) service proposed for SGEMS. Basically, the model demonstrates a technique for allocating resources from multiple resource pool (domain servers) in the cloud Data Center, while showing how the effect of resource allocation using virtualization can affect server performance. From the literature survey, little research has been done in this area, as such; this work seeks to use its findings to consolidate on the network performance of the SGEMS whose support platform is DCCN. Various mathematical models were derived to address performance constraints in a cloud based resource pool while using a Branch-and-Bound heuristic procedure to find good metric values for the evaluation study. From MATLAB Simevent environment, it was shown that that the server that have higher resource pool (virtualized type) offered an average delay of 21.76%, with average service availability 64.42%, hence having more job completion times. On the other hand, the server with lower resource pool (legacy type) offered an average delay of 78.24%, with average service availability of 35.58%, hence having lower job completion times. Consequently, from the experimental results, it was concluded that with virtualization in a cloud DataCenter servers, larger pool sizes in server resources results in a lower average wait time even with a higher request rate and longer duration of resource use while offering fault tolerance optimally. [