摘要
面对当前网络中流量的增长、业务种类的增多,SDN中多数的路由算法只支持一种QoS参数,没有兼顾对系统调度服务公平性的考虑,然而多参数限制的QoS明显是NP难问题,该问题用普通的路由算法难以解决,引进蚁群算法,在蚁群算法的基础上,将链路的时延、分组丢失率引入蚁群算法中,作为算法选择路径的依据,提出一种新的路由算法。该算法在对不同业务属性的数据流分类的基础上,根据网络的实时状况,为不同业务属性的数据流选择合适的路径,对网络中的数据流进行多路径传输。仿真实验表明,该算法能有效地降低数据流的时延、分组丢失率。
Currently, most of the routing algorithms support only one QoS parameter, and do not consider the fairness of system scheduling service. Obviously, multi-parameters constrains problem is NP hard problem, it is impossible to solve this kind of problem by traditional routing algorithms. The routing thought from ant colony algorithm(ACO) was introduced. On the basis of the ACO, and combined with the business attributes of traffic flow,a new algorithm was put forward which introduced the delay and packet loss into the ACO, as a dependence of routing. In new algorithm, which according to the real-time status of network, the suitable routing was chosen for different types traffic and forward by multi-path base on the classification of traffic flows. The experiment shows that this algorithm can decrease the delay and packet loss effectively.
出处
《电信科学》
北大核心
2016年第11期10-15,共6页
Telecommunications Science