Dear all, On the #DENOG IRC channel I was asked for current stats on the number of RPKI-invalid IRR route/route6 objects in various databases as follow-up to a talk at RIPE81 [0]. I figured I should share this with the WG too. Below is a table with today's stats of number of invalid route/route6 objects when one applies the RFC 6811 origin validation algorithm with as input prefix value in the "route:" attribute and the origin AS in the "origin:" attribute. invalids invalids AFRINIC: ipv4: 359 - ipv6: 12 - authoritive ALTDB: ipv4: 1 - ipv6: 191 - note 4 APNIC: ipv4: 21861 - ipv6: 1880 - authoritive ARIN: ipv4: 814 - ipv6: 65 - authoritive BBOI: ipv4: 44 - ipv6: 1 BELL: ipv4: 322 - ipv6: 0 JPIRR: ipv4: 95 - ipv6: 4 LACNIC: ipv4: 0 - ipv6: 0 - authoritive (note 3) LEVEL3: ipv4: 12925 - ipv6: 182 NTTCOM: ipv4: 65513 - ipv6: 730 RADB: ipv4: 208901 - ipv6: 12829 RGNET: ipv4: 2 - ipv6: 0 RIPE: ipv4: 28390 - ipv6: 3518 - authoritive RIPE-NONAUTH: ipv4: 5 - ipv6: 0 - note 5 TC: ipv4: 0 - ipv6: 0 - note 2 Some notes on the above table: 1) ARIN-NONAUTH is not listed, ARIN deprecated this IRR source a month ago [2]. 2) TC achieved a perfect 0/0 score by using the IRRd v4 RPKI integration [3]. 3) LACNIC's IRR service is an information proxy for RPKI ROAs valid under the LACNIC Trust Anchor. This by definition means that all IRR objects in the LACNIC IRR database are RPKI-valid. 4) ALTDB periodically runs a script to delete RPKI-invalid objects 5) RIPE-NONAUTH imposes a two week delay before deleting RPKI-invalid objects, so the 5 IPv4 objects currently marked as invalid with disappear in the next few days, unless the covering RPKI ROAs are withdrawn before the timer expires. The stats are generated by downloading the IRR database dump for each source and running a simple python script [1]. Kind regards, Job [0]: https://ripe81.ripe.net/presentations/59-IRRd-RIPE812.pdf [1]: https://github.com/job/irr-nonauth-cleanup [2]: https://www.arin.net/announcements/20220128-irr/ [3]: https://irrd.readthedocs.io/en/stable/admins/rpki/