Abstract: The high dynamics of the satellite network topology and the large link transmission delay have brought about problems such as frequent link switching and high packet loss rate, which leads to local congestion in the network and cannot meet business QoS requirements. In order to solve the above problems, a multi-QoS constraints routing algorithm based on dynamic resource graph of virtual nodes (DRGVN-QR) of SDN architecture is proposed. First of all, in order to better characterize the dynamics of node switching and resources, combined with the network topology of virtual nodes, the dynamic resource graph of virtual node model is established, and a multi-objective optimization model based on the minimum path cost is established. This model considers the node's switching status, cache, and link remaining bandwidth, delay and other information; then, use the ant colony algorithm to solve the objective function, and concurrently find the optimal path set within a period of time for each connection request, in order to improve the path quality and algorithm performance, the value of the pheromone volatilization coefficient is discussed; finally, in order to adapt to the time-varying nature of the satellite network, a power weighting formula is designed to find the optimal path within a period of time. The simulation results show that the DRGVN-QR algorithm can effectively relieve network congestion and improve network QoS. Compared with other algorithms, this algorithm reduces the average end-to-end delay, network packet loss rate and delay jitter.

Key words: satellite network, SDN, dynamic resource graph, multi-objective optimization, ant colony algorithm