Cost and energy optimisation of cloud data centres through dual VM modes - activation and passivation

Abstract

To support cloud computing, new generation data centres are dependant on virtualisation. Server consolidation and live migration of virtual machine are two important techniques to attain energy savings and load balancing. Energy efficiency, performance and cost incurred are different tradeoff parameters to be considered. We propose a queueing-based analytical model for performance of VMs under dual operational modes, that is, activation and passivation. Active VMs with higher computational power are used when the load on the system is high, otherwise, passive VMs are used. We achieve high-availability, load-balancing, fail-over, less energy consumption as well as cost optimisation through clustering of both active as well as passive VMs. To avoid congestion and to adhere the SLA under fluctuating workload, the system can easily switch from one mode to the other under certain conditions. Numerical results are depicted to prove the applicability of the proposed model for the data centre management.