High-Risk LEO Satellite Network Path Detection Based on Spatial and Temporal Delay Anomaly Analysis

The gradual deployment of Low-Earth Orbit(LEO)mega constellations with inter-satellite links(ISLs)promises ubiquitous,low-latency,and high-throughput satellite network services.However,networked LEO satellites with ISLs are also at risk of routing attacks such as hijacking.Existing de-fenses against route hijacking in terrestrial networks can hardly work for the LEO satellite network due to its high spatiotemporal dynamics.To deal with it,we propose RPD,a high-risk routing path detection method for LEO mega-constellation networks.RPD detects abnormal high-risk LEO network paths by checking the consistency between the path delay and the geographical distance.This is efficiently achieved by combining in-band measurements and out-of-band statistical processing to detect the anomaly of the clus-tering feature in the reference delay matrix.RPD avoids the recalculation of the header cryptographic marks when the handover occurs,thus greatly reduc-ing the cost and improving the performance of high-risk path detection.Experiments showed that the pro-posed RPD mechanism achieves an average detection accuracy of 91.64％under normal network conditions,and maintain about 89％even when congestion oc-curs in multiple areas of the network and measurement noise is considered.In addition,RPD does not require any cryptographic operation on the intermediate node,only minimal communication cost with excellent scal-ability and deployability.