Emperor Penguins’ Behavior-based Virtual Machine Placement Algorithm for avoiding heat islands in Cloud Computing Environment

In cloud computing environment, a single server accommodates multiple virtual machines for resource efficiency. It is necessary to place virtual machines on servers according to the energy and resource efficient policy. This paper proposes a virtual machine placement (VMP) algorithm based on the emperor penguins’ behavior, which attempts to minimize loss of penguins’ body temperature.
Penguins with the greatest heat loss continuously move in their group against strong external winds.
The proposed algorithm attempts to place virtual machines on servers for minimizing power consumption, and periodically migrate virtual machines on the server, which possibly occurs a heat island, to a server with low temperature. This paper proposes server-based and rack-based VMP algorithms. Simulation results show that the two proposed algorithms consumed server power similar to the existing algorithm while virtual machines were placed in a mostly distributed way. In case of applying the rack-based VMP, the difference in power consumption among servers is the smallest so that this algorithm is considered to be the most suitable for preventing the heat island.