Local-state routing in satellite constellation networks from the perspective of complex network

Routing algorithms in satellite constellation networks usually make use of the local state information to adapt to the topology and traffic dynamics, since it's difficult to obtain the global states in time due to the spatial large-scale feature of constellation networks. Furthermore, they use different range of local states and give these states distinct weights. However, the behind design criterion is ambiguous and often based on experience. This paper discusses the problem from the perspective of complex network. A universal local-state routing model with tunable parameters is presented to generalize the common characteristics of local-state routing algorithms for satellite constellation networks. Based on this, the impacts of local-state routing algorithms on performance and the correlation between routing and traffic dynamics are analyzed in detail. Among them, the tunable parameters, the congestion propagation process, the critical packet sending rate, and the network robustness are discussed respectively. Experimental results show that routing algorithms can achieve a satisfactory performance by maintaining a limited state awareness capability and obtaining the states in a range below the average path length. This provides a valuable design basis for routing algorithms in satellite constellation networks.