Optimization of the Number of Servers in Cloud Computing Centers

Abstract

To improve the service quality and save the system cost of the cloud computing center, this paper studies the joint optimization problem of energy consumption and performance of cloud computing centers with a batch Markovian arrival process. The system has multiple parallel processors and the processing time of each processor follows phase type distribution. The system has finite buffer. We construct the system as a BMAP/PH/N/M queueing system and analyze the performance of the system based on queuing theory. An optimization model is established by combing the energy consumption and system performance measures. The optimal number of servers is analyzed. Some managerial insights are given by numerical analysis.