BGP Failure Propagation (BGP-FP) for Enhancing Control-Plane Convergence

This document specifies BGP Failure Propagation (BGP-FP), an infrastructure and protocol that improves inter-domain routing convergence by accelerating the removal of stale (invalid) routes. BGP-FP uses (1) an Agent deployed per Autonomous System (AS) to detect inter-AS reachability changes and to configure local routers, (2) a logically centralized Repository to store and selectively forward AS reachability state, and (3) BGP Large Communities as a "route freshness" marker. Agents validate and apply Repository updates to filter routes that traverse AS pairs whose reachability has been lost or that violate the originating AS's forwarding intent, reducing route-flap propagation in the control plane. This document clarifies that "AS reachability" refers to the reachability between two ASes, not to the state of individual physical or logical links within an AS. If multiple links exist between two ASes, the failure of a single link that does not break overall AS-to-AS reachability does not trigger the BGP-FP mechanism. A new Repository deployment model is introduced, suggesting that the Repository be operated by a newly established organization composed of Tier-1 ASes and Regional Internet Registries (RIRs), using a distributed deployment with Byzantine fault-tolerant consensus and rotating leadership.