Topology Semantic Routing in Multi-Layer Satellite Networks

Mega-constellations introduce frequent and predictable topological changes due to orbital motion, link dynamics, and weather-affected feeder availability. Classical routing approaches relying on global re-signaling are unable to maintain stable connectivity under such dynamics, leading to increased overhead, delayed convergence, and degraded delivery. In this article, we introduce the topology semantic routing concept that enables space nodes to adapt their forwarding behavior based on local and momentary topology interpretation. Routing protocol signaling is deferred or avoided for transient or expected changes, and path selection is adjusted using time-aligned information or localized failure detection. The mechanism integrates with distributed and centralized frameworks and is suited for homomorphic and cyclic topologies, frequently occurring in mega-constellations. Furthermore, the results of a simulation-based evaluation under orbital shifts, feeder handovers, and link degradation show that time- and event-driven redirect strategies maintain delivery performance and reduce signaling overhead, confirming applicability in dynamic non-terrestrial infrastructures.