基于改进的粒子群优化算法的SDN负载均衡研究

Research on Load Balance for Software Defined Network Based on Improved Particle Swarm Optimization Algorithm

链路的负载均衡是数据中心网络中需要考虑的核心问题之一,当前众多学者基于SDN解决数据中心大象流负载均衡的问题。通过设计并改进一种多商品流的粒子群优化算法并用于求解负载均衡问题。首先构建基于路径长度与链路利用率的大象流分布的算法目标函数,接着提出一种改进的IPSO算法,该算法很好地避免粒子群算法在搜索后期易陷入局部最优的困境,最后基于构建的Floodlight和Mininet环境的SDN数据中心实验平台进行仿真实验,实验结果表明IPSO算法具有更高的可靠性,获得更好的链路负载均衡效果。

Link load balancing is one of the key problems that need to be considered in the data center network, many scholars solve the elephant flow load balancing problem based on the SDN data center. Through designing a kind of multi-commodity flow and the particle swarm optimization algorithm is used to solve the load balancing problem. First builds flow distribution algorithm based on path length and the link utilization, the objective function of the elephant, and then puts forward an Improved Particles Swarm Optimization （IPSO） algorithm, the algorithm is very good to avoid the Particle Swarm Optimization （PSO） algorithm in the search in the late fall into the predicament of the local optimum easily, based on the building Floodlight and Mininet environment SDN data center experiment platform for simulation experiment, the experimental results show that IPSO algorithm has higher reliability, and obtained a better link load balancing effect.