An Energy Efficient Load Balancing Algorithm Based on the Active Time of Cores

Server cluster systems are widely used to realize scalable and high performance computing systems with virtual machine technologies. A large amount of electric energy is consumed in a server cluster system since a server cluster system is composed of large number of servers and multiple servers consume electric energy to perform application processes on multiple virtual machines. In order to design and implement an energy-efficient server cluster system, it is necessary to realize energy-efficient load balancing algorithms. In this paper, the active time-based (ATB) algorithm is proposed to select a virtual machine for each request process so that the total electric energy of a server cluster to perform computation type application processes can be reduced. In the ATB algorithm, it is not necessary to collect a state of every process on every virtual machine to estimate the electric energy of each server. We evaluate the ATB algorithm in terms of the total electric energy of a server cluster compared with the energy consumption laxity based (ECLB) and basic round-robin (RR) algorithms.