A study of resource management for fault-tolerant and energy efficient cloud datacenter

Abstract

In cloud computing datacenters, the reliability and energy consumption have been studied as main challenges to achieve the reputation of cloud service users and the cost efficiency. To overcome the system fault of the datacenter, VM request load has to be distributed on multiple hosts to minimize the effect to the running cloud applications. Moreover, Dynamic Right Sizing (DRS) which adjusts the number of active hosts and sleep hosts in order to reduce the energy consumption in view of the resource usage cost. To do this, we propose the resource management scheme based on the portfolio diversification which has been studied in economics. The proposed scheme is able to reduce the fault of application significantly by finding the near Pareto optimal solution through Simulated Annealing approach We show the efficiency of our proposed scheme through the simple analytical results.