GraphRoute-Transfer: Topology-Generalizable Routing Convergence Optimization via Graph Reinforcement Learning
Journal of Computing and Electronic Information Management · 20.07.2026 · 22 (1) · 12-18
DOI: 10.54097/h0c18694 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Computing and Electronic Information Management, online erschienen am 20.07.2026. Verfasst von Yuto Nakamura. Die Arbeit nennt 30 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 4.0).
Abstract (Originalsprache)
Fast and stable routing convergence is critical in large IP networks, and the interior-gateway-protocol (IGP) timers that govern failure detection (Hello/Dead intervals) expose a fundamental tension: aggressive timers detect failures quickly but inflate control overhead and trigger route flaps, whereas conservative timers are stable but slow. Recent work such as DRL-Adapt has shown that deep reinforcement learning can tune these timers better than static defaults, but it operates on a flat, globally-aggregated network state and emits a single network-wide timer, so it can neither exploit the spatial heterogeneity of real topologies nor transfer architecturally across networks of different size. We propose GraphRoute-Transfer, a graph-neural-network policy that assigns per-node timers from local structural features and is by construction permutation- and size-invariant. Because control-plane fragility and failure criticality are spatially heterogeneous, the cost-minimizing timer assignment varies across the graph; our policy learns this mapping and applies it zero-shot to unseen topologies of arbitrary size. Training is guided by a coordinate-descent search oracle on a convergence-cost objective, so the expensive per-topology optimization is amortized into a sub-millisecond inference. On 231 real topologies from the Internet Topology Zoo, GraphRoute-Transfer reduces mean convergence time by 37.3% relative to the OSPF default and to a flat DRL baseline, lowers the composite convergence-cost objective by 16.4% over the flat baseline, and attains 1.317 cost—within 0.3% of the search oracle—while running about 8,160× faster than the search. Crucially, a policy trained only on networks with ≤70 nodes maintains its gains on unseen networks up to 140 nodes, whereas the flat baseline degenerates to a global constant that cannot adapt.
Zitieren
APA
Nakamura, Y. (2026). GraphRoute-Transfer: Topology-Generalizable Routing Convergence Optimization via Graph Reinforcement Learning. Journal of Computing and Electronic Information Management, 22(1), 12-18. https://doi.org/10.54097/h0c18694
Vancouver
Nakamura Y. GraphRoute-Transfer: Topology-Generalizable Routing Convergence Optimization via Graph Reinforcement Learning. Journal of Computing and Electronic Information Management. 2026;22(1):12-18. doi:10.54097/h0c18694
BibTeX
@article{Nakamura2026,
author = {Nakamura, Yuto},
title = {GraphRoute-Transfer: Topology-Generalizable Routing Convergence Optimization via Graph Reinforcement Learning},
journal = {Journal of Computing and Electronic Information Management},
year = {2026},
volume = {22},
number = {1},
pages = {12--18},
doi = {10.54097/h0c18694}
}
EUVIMED