Due to the dynamic changes of network state or environment, load balancing faces complexity challenges and it is difficult to accurately predict future load conditions. Therefore, a dynamic traffic load balancing algorithm considering the priority of nodes is proposed. The network analysis tool is used to obtain the load of node traffic data received by sFlow technology, and the integration process is carried out to obtain the basic data set of node load. According to the node load on the data set, the node priority method is used to select the node that can be preferentially allocated more traffic. An ecological predator-prey model based on ecological difference equation is established. The selected nodes are taken as the input of the model, and the dynamic balance of network traffic load is realized through the iterative update of the model. The experimental results show that the proposed traffic load dynamic balancing algorithm can improve network throughput and CPU utilization, and has better practical application effect.

communication traffic load, node priority, sFlow technology, dynamic balancing algorithm, ecological difference equation