Dynamic load balancing strategy for cloud computing with ant colony optimization

Abstract

How to distribute and coordinate tasks in cloud computing is a challenging issue, in order to get optimal resource utilization and avoid overload. In this paper, we present a novel approach on load balancing via ant colony optimization (ACO), for balancing the workload in a cloud computing platform dynamically. Two strategies, forward-backward ant mechanism and max-min rules, are introduced to quickly find out the candidate nodes for load balancing. We formulate pheromone initialization and pheromone update according to physical resources under the cloud computing environment, including pheromone evaporation, incentive, and punishment rules, etc. Combined with task execution prediction, we define the moving probability of ants in two ways, that is, whether the forward ant meets the backward ant, or not, in the neighbor node, with the aim of accelerating searching processes. Simulations illustrate that the proposed strategy can not only provide dynamic load balancing for cloud computing with less searching time, but can also get high network performance under medium and heavily loaded contexts.