Stableroute Algorithm For Dynamic Topology Of LEO Satellite Network

With the rapid development of global LEO satellite networks and related industries,LEO satellite constellation is becoming a viable means for Internet access,reflecting their significant role in various aspects.In the new era,the LEO satellite network represents a strategic opportunity.However,its topology changes frequently as satellites move into or out of polar region(link down or up).This may cause FTE updates and thus path changes between two satellites.A path change during transmission may result in out-of-order packet delivery or packet losses.The current scale of satellites is gradually expanding,with Musk’s "Starlink" and China’s "Integration of Space and Earth" indicating that the scale of satellite networks will be rapidly upgraded.Therefore,the routing problem of such satellite networks cannot be ignored.In the research process,we found some path changes are inevitable,while some are avoidable.Dijkstra’s algorithm is a popular choice among the routing protocols proposed for LEO satellite networks.We observe that many FTE updates by Dijkstra’s algorithm are avoidable.Motivated by this,we propose StableRoute,which stabilises routing paths from different perspectives.StableRoute Local(SR_L)leverages equal-cost shortest paths and stays with the current one if it is still valid,regardless the choice of Dijkstra’s algorithm.StableRoute K-Short(SR_K)allows a path longer than the shortest path.As long as the current path is valid and its cost is not higher than the least cost by a predefined value,the current path stays.StableRoute Global(SR_G)takes advantage of the predictable satellite trajectories and topology variations to work out the FTE update rules that minimise the number of FTE updates.The evaluation shows that SR_L,SR_K and SR_G outperform Dijkstra’s algorithm,and reduce the number of FTE updates by 43.40%,44.02%,and 45.86%,respectively,for a 40 × 40 constellation.