Multi-Layer Routing Domain Protection at Hyperscale: From Route Leak to Automated Enforcement

Operating a global backbone connecting dozens of data centers requires strict enforcement of which prefixes are allowed to cross domain boundaries. A single misconfiguration — a wrong BGP community tag or an unregistered prefix — can cascade into widespread service disruption across multiple network domains.
This talk presents our journey building a multi-layered defence system against routing domain violations at scale. After a major routing incident exposed gaps in our boundary protection, we developed three independent, complementary layers:
(1) ID Communities — a tagging system assigning each prefix a unique identity that defines its origination and propagation scope across domain boundaries.
(2) Route Registry — a source of truth defining which prefixes are explicitly permitted at each boundary, deployed as inbound prefix-lists with a freshness SLA.
(3) Control Plane Health Monitoring — an automated system that continuously validates the routing state against intent, detecting community mismatches and unregistered prefixes before they become incidents.
We share how this system exposed pervasive violations across production, prefixes crossing boundaries without correct identity, and how we eliminated nearly all of them within months through systematic remediation. Automated enforcement now prevents recurrence.The talk concludes with lessons learned on making routing domains self-protecting at hyperscale.
Relevant to operators managing multi-domain networks, backbone/DC boundary protection, and anyone interested in intent-based routing verification.