SDN 中基于负载均衡的最优路径分配算法研究

摘要

针对传统网络控制与转发的紧耦合，其路径分配算法的全局性差、实时性不足，根据软件定义网络的集中控制，提出一种基于马尔可夫链负载均衡（Markov chain-load balancing，MC-LB）算法。该算法使用马尔可夫链的转移概率计算链路重要性，并同时对业务流请求量和网络负载均衡率两个目标进行优化，获得多请求下的最优路径分配。通过大量实验对比表明，该算法优于传统 SPF（shortest path first）算法，最大可以增加网络负载均衡率30％，提升网络业务流请求接受率20％，达到了负载均衡效果，提高了网络的性能。%To solve the problem that path allocation algorithm has poor globality and lack of real-time in the tightly couples of traditional network control and forwarding,the paper used centralized control of software defined networking (SDN)to propose the MC-LB algorithm.The method calculated link importance by transition probability of Markov chain,and optimized the business flow request quantity and the load balance rate of network simultaneously to get optimal path in multi-requests.Large experiment results comparing analysis prove that the algorithm is superior to the traditional SPF (shortest path first)algorithm which can increase the 30% network load balance rate and 20% business flow request acceptance rate in some cases.The pro-posed algorithm achieves the effect of load balancing and improves the performance of network.