Modeling and Analysis of Cloud Computing Network using M/G/k/PS Queues

The cloud computing are gaining more and more popularity in business, enterprises and research areas. Flexible and transparent provision of services on-demand, height virtualization level, seamless interconnection between different network nodes are the reasons of good performance and scalability of cloud computing network. However cloud computing technology has not universal requirements and solutions for network deployment and implementation. The lack of standards leads to performance and functionality reduction. The analysis of functionality and performance of cloud computing solutions give ability find bottlenecks and improve services provision processes. The difference types of cloud computing architecture are analyzed in the paper. The different approaches to the performance and functionality analysis, such as real measurements, imitation and analytical modeling are observed. The analytical model of cloud network based on queuing theory is proposed. The selecting of queuing model parameters depended on the technology platform, the model of cloud services (SaaS, PaaS, IaaS), the type of applications. Proposed queuing model allows analyzing the main network characteristics affected on performance and services provision effectiveness.