ROV impact measurement study
We have a ROA that authorises AS211321 to announce 2a04:b905::/32-32 and 18.104.22.168/23-23. We announce 2a04:b905::/32 and 22.214.171.124/23 from Vultr in Sydney, and 2a04:b905::/33 and 126.96.36.199/24 from ColoClue in Amsterdam. If everyone would do ROV, then the traffic inside 2a04:b905::/33 or 188.8.131.52/24 (e.g. 2a04:b905::2) should go to Sydney. Our hypothesis is that if one link in the chain is missing, that the traffic will not end up in the intended location.
To measure this we have set up two RPKI publication points, parent.rov.koenvanhove.nl (2a04:b905:8000::1 and 184.108.40.206) and child.rov.koenvanhove.nl (2a04:b905::2 and 220.127.116.11). All traffic will go to Sydney for the parent (because that's the only place it's announced), and we measure where the traffic for the child ends up.
To verify if a network is doing ROV and dropping invalids, we have also set up another publication point called invalid.rov.koenvanhove.nl (2001:ddb::1 and 18.104.22.168), which are announced from Vultr in Sydney, and have an AS0 ROA (22.214.171.124/24-24 and 2001:ddb::/48-48) associated with them (and no less specific announcement exists). Regardless of what the upstream does, a network dropping invalids should not send any traffic to these endpoints to its upstreams.
This research is a collaboration between the University of Twente, RIPE NCC, and NLnet Labs. The article can be found on RIPE Labs and the NLnet Labs blog.
Information regarding your current connection
You might be interested where your own traffic ends up. Once the page has loaded, your browser should make three requests to the three endpoints. The results are shown below.
|Connects to parent||Child location||Can visit invalid|