Optimized Dynamic Load Balance Method Based on Ant Colony Optimization Algorithm

Oriented to dynamic load balance of high concurrent cluster, an Optimized Dynamic Load Balance Method (ODLBM) based on Ant Colony Optimization (ACO) is proposed in this paper. Compared to traditional static load balance approach, ODLBM tends to resolve complex dynamic load balance problem with dynamic cluster and heterogeneous tasks. First, the dynamic load balance problem is transformed to the optimization problem based on ACO. After that, the optimized calculation of distance via comprehensive indicator of weights, and the optimized update of pheromone are introduced into the ACO model to improve the performance of basic ACO model in complex dynamic load balance scenario. Finally, the efficiency of ODLBM is validated by the experiment based on JMeter simulation, compared to the classical Weighted Round Robin (WRR) method, and Fastest Response Time (FRT) method.

I. Introduction

With the development of internet industry, internet users are increasing exponentially. In terms of large scale of internet users and business, the biggest challenge from stability and performance of internet service is to relieve the heavy load of servers, which may break the system. Thus, in this paper, an efficient strategy oriented to high concurrent access of internet is proposed to relieve the heavy loads of servers and improve the robustness of system. In this strategy, Optimized Dynamic Load Balance Model (ODLBM) based on Ant Colony Optimization (ACO) Algorithm utilizes high concurrent cluster technology to increase the concurrent processing capability of servers.