一种适用于星地网络的流量负载均衡算法

在星地网络的背景下，为了降低由于地面网络链路负载过高导致的传播时延，提出了一种适用于星地网络的负载均衡算法。算法基于网络的链路容量设置一个链路资源利用率作为阈值，当发现路径出现过载的时候分成两步进行路径的预规划：第一步，运用蚁群算法为链路过载的数据流选择其他的路径，以满足地面网内部的负载均衡；第二步，如果仍出现链路超阈值的情况，通过卸载延迟容忍数据流来执行地面与卫星网络之间的负载均衡。仿真结果表明，设定自适应的过载阈值可以有效降低链路的拥塞，与现有算法相比该算法不仅可以更加均匀地分配地面网络中的流量，而且借助卫星网络使地面网络的平均时延降低了18.3%，提高了网络的服务质量。

A traffic load balancing algorithm suitable for satellite-terrestrial networks

In the context of the satellite-terrestrial networks,a load balancing algorithm suitable for satellite-terrestrial networks is proposed to reduce the propagation delay caused by the excessive load of the terrestrial network link.The algorithm sets a link resource utilization rate as a threshold based on the link capacity of the network.When the path is found to be overloaded,it will be divided into two steps to pre-plan the path.In the first step,the ant colony algorithm is used to select other paths for the data flow the link of which is overloaded to perform load balancing within the terrestrial network.In the second step,if the situation that the link is still higher than a predefined threshold exists,the load balance between the terrestrial and the satellite network is performed by offloading the delay-tolerant data flows.Simulation results show that setting an adaptive overload threshold can effectively reduce link congestion.This algorithm can not only distribute the traffic in the terrestrial network more evenly,but also reduce the average delay of the terrestrial network by 18.3% through the satellite network when compared with existing load balancing algorithms.