New Pre-Rerouting Strategy Safeguards Business Continuity for Low-Earth-Orbit Satellite Networks

Low Earth orbit satellite networks could soon reroute communications traffic before a connection disappears, rather than waiting for an interruption to occur. A study published in Space: Science & Technology presents a pre-rerouting strategy designed to preserve service continuity as satellites move through orbit and their inter-satellite links repeatedly switch on and off. The approach extends the widely used Open Shortest Path First routing protocol so that satellite routers can consider not only the distance and bandwidth of a link, but also how long that link is expected to remain usable. In tests, the system switched traffic to a backup route before a predictable link interruption, reducing packet loss from 7% to 1% in a hardware-in-the-loop simulation.

The need for such a system is becoming increasingly urgent as low Earth orbit constellations expand. LEO satellites can provide broad coverage and connect regions that lack terrestrial infrastructure, making them important components of emerging space–air–ground integrated networks and future 6G communications. Yet these networks are fundamentally different from conventional fixed communication systems. Satellites are constantly moving relative to one another, and links between them may be available only for limited periods. Some connections disappear according to predictable orbital patterns, while others can be affected by changing network conditions. A routing system that treats every active link as permanently available may select a path that becomes unusable before a data session is finished.

Traditional OSPF routing generally selects paths according to a metric associated with link cost, often favoring routes with fewer hops or lower transmission distance. That strategy works well in relatively stable terrestrial networks, but it does not adequately capture the time-dependent nature of satellite connectivity. A short route may be attractive from a distance perspective while offering only a few minutes of remaining availability. If a video stream, control signal, or other service requires a longer transmission period, the route can fail in the middle of the session. Fast reroute mechanisms can respond to sudden failures, but they typically require reserved backup resources and may not protect every node or connection across a large constellation.

To address these limitations, the researchers added two fields to the routing information distributed throughout the network. The first records the termination time of a link’s validity, allowing routers to estimate how long the connection will remain available. The second describes available bandwidth, enabling the routing process to reject paths that cannot support the intended service. These additions extend the information carried by a link-state advertisement, or LSA. By exchanging the updated information through a Link Management Protocol, routers can maintain a synchronized view of the network’s changing conditions. In practical terms, every participating node can make routing decisions based on both the remaining lifetime and the capacity of each available connection.

The study also changes how routes are evaluated. Instead of relying only on a conventional distance-based metric, the proposed system combines the traditional cost with link effective duration through a weighted calculation. Adjusting the weighting factor allows the network to prioritize different operational goals. A system carrying delay-sensitive traffic might favor a shorter route, while a long-duration service could place greater emphasis on route stability. The researchers use a Constrained Shortest Path First algorithm to select paths that satisfy both bandwidth and time requirements. In an illustrative example, a route with fewer hops becomes unusable after 200 seconds and therefore cannot support a service expected to last 800 seconds. A slightly longer route, however, remains available for up to 1,000 seconds and is selected by the time-aware algorithm.

The core innovation is a trigger mechanism that initiates rerouting before a link failure. Once a label switched path, or LSP, has been established, the source node continuously monitors its remaining effective lifetime. If that remaining time becomes shorter than the service’s remaining transmission duration and also falls below a predefined threshold, the source node begins preparing a new route. It calculates a backup path using the latest bandwidth and duration information, reserves resources along that path, and establishes a new LSP through a constraint-based routed label distribution protocol. Traffic is then transferred to the new path, after which the resources used by the old path are released. This sequence is intended to prevent the interruption that would occur if rerouting began only after the original link had already vanished.

The researchers tested the mechanism on an embedded hardware platform combining the VxWorks operating system with field-programmable gate array hardware. In the experimental topology, dynamic satellite-link behavior was represented by connecting and disconnecting ports between control nodes. The original LSP was deliberately established through a port with a shorter remaining validity period. When the pre-rerouting condition was met, the source node identified a route using a port expected to remain available longer, established the new path, and moved the service before the original connection ended. Ping traffic continued normally during the transition, and the researchers reported no packet loss in this functional verification. The result indicates that the process can be executed autonomously by network hardware rather than relying on manual intervention.

A second evaluation used the EXata hardware-in-the-loop simulation platform to compare the proposed method with conventional OSPF rerouting. The simulated network provided two alternative routes between a source and destination. The primary route was configured to remain available for 100 seconds, while the service continued for 500 seconds. In the conventional scenario, the link failed at the 100-second mark, and traffic resumed only after an interruption of approximately 20 seconds. Out of 1,000 UDP packets transmitted by the source, 930 reached the destination, corresponding to a reported packet loss rate of 7%. When pre-rerouting was enabled, the service moved to the backup path before the predicted failure. The destination received 990 packets, producing a 1% loss rate and an improvement of approximately six percentage points.

The results show why predictive routing may be particularly valuable for satellite networks, where many topology changes are not random failures but consequences of orbital motion. By treating time as a routing constraint, the proposed system can avoid paths that are technically available but operationally unsuitable for the duration of a service. It may help reduce disruptions in broadband access, remote sensing data transfers, satellite control, and other applications that cannot tolerate a lengthy reconnection delay. The researchers acknowledge that the backup route selected before an interruption may use more hops or introduce greater transmission delay than the original path. They suggest that future systems could recompute a globally optimal route after the topology changes, combining uninterrupted service with later performance optimization. For rapidly expanding LEO constellations, the strategy offers a practical way to turn predictable orbital motion from a source of disruption into information that can be used for advance planning.

News Publication Date: 1-Jul-2026

Web References: https://doi.org/10.34133/space.0420

References: Space: Science & Technology, DOI: 10.34133/space.0420