Resource Public Key Infrastructure (RPKI is being adopted rapidly: the share of routed prefixes marked RPKI-valid climbed from 54.0% in September 2025 to 70.3% in September 2026. RPKI lets address holders publish signed Route Origin Authorizations (ROAs) that name the Autonomous Systems allowed to announce their blocks, and networks that validate RPKI can drop invalid announcements. The year’s surge was driven overwhelmingly by China (valid IPv4 coverage rising from 3.7% to 83.0%), with big gains also in Egypt, Kazakhstan, Mexico and South Africa. IPv6 adoption has outpaced IPv4 (roughly 79% vs 68% RPKI-valid). Invalid routes now rarely propagate - typical invalid prefixes are seen by only two public vantage points vs ~596 for valid ones - so raw counts of invalid prefixes can exaggerate risk unless weighted by address space and visibility.
RPKI is proving effective but not foolproof. It only validates origin AS, not path, and permissive ROA maxLength settings create “open windows” for more-specific hijacks: a recent Hetzner case showed a forged /24 could be RPKI-valid because its parent /16 ROA allowed /24s. About 16.2% of covered IPv4 routes expose such windows (20.9% for IPv6), with wide operator variation. Coverage can also lapse when ROAs disappear (examples: Mauritius, SEACOM, Saudi Telecom, Orange Maroc). Finally, some measurement methods falsely report record failures when repository links are stale; correct validation walks from trust anchors and yields different coverage numbers.
Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.