the 3 day version is even more entertaining - about half way through you start to see that there are apparently two dominant calendar systems at work

Some of the noise might just be a quirk of the visualisation approach. If anything, to me it shows that most of RPKI’s automated systems favor issuing somewhere at the top of the hour! #cronjobs

Because at say 14:00 there *no longer* are many current manifests around that were issued earlier at say 01:00. The animation is a “viewport” of the internal state of a validator throughout the day

Happy to come present in exchange for a flight ticket! :-)

It went well at my last employer

if you’re using bgpq4, the filters it emits no longer contain any information derived from RPKI-invalid IRR objects, but conversely _do_ use synthesised IRR objects derived from RPKI ROAs (i take credit for bridging the two domains). In some ways the migration from IRR to RPKI already happened

Even so, what’s the issue with RADB? That database doesn’t contain RPKI-invalid IRR route objects (since they deployed IRRDv4). Considering to not use RADB goes to show that maybe we don’t really really need IRR-based filtering that bad at all?

don't under-estimate the significance of IRR data being unauthenticated plain-text without any cryptographic assurances!

Those spammers can create IRR objects in altdb/level3/radb/etc just like that! What you want is BGPsec! You want more RPKI :-)

It’s allowed, unless there are conflicting signals (such as ROAs, or maxpfx limit, peer lock filter, ASPAs, or OTC/RFC9234 appears in wrong context). If none of those things are cause for rejection, the route is innocuous. That’s not a less secure paradigm compared to utterly insecure IRR data

Nah… you’ll come around 8-)

I repeat: the objective is to disallow invalid routes (things can only become invalid iff an ASPA or ROA exists), not to require all routes to be valid

This is the key bit: I’m not proposing to cut folks off! It’s just that when you don’t create ROAs or ASPAs you enjoy less protection. End user decides how much risk they want to expose themselves too