> You are twisting the fact when you are mixing hosted-rpiki and > rpki-repositories as the same thing.
Please feel free to point to the place in any message when I actually said they are the same thing. They are part of the same system, not the same thing --and I don't think I ever claimed otherwise. > No, I am saying that rpki hosted (the web interface) is not critical > for RPKI. Depends on who's ox is being gored, I suppose. For the provider/RIR, it's not so crucial. For the customer, it seems like it might be (depending on the customer, etc.). I would think that if I were losing x million of dollars per minute because my site is unreachable, I might consider being able to reach the services required to bring my site back onto the web to be somewhat important... Along with the time required to propagate information through that system, etc. > A bit, you are (in my understanding) are still mixing hosted-rpiki and > rpki-repositories. ;-) No... There is an entire ecosystem here, not just one "thing." There is the certificate structure, four or five kinds of certificates themselves, the data replication service (possibly multiples), a set of protocols for getting the information between the replicating server and the router, and now, from what I understand, a hosted service as well... The point Eric was making is that each piece of the system adds complexity. The counter was, "what complexity?" My answer is --each and every piece adds complexity, no matter whether each particular piece is considered "crucial," etc. More moving parts == more complexity Sometimes internal complexity is a good tradeoff for external simplicity, but that doesn't make the complexity issues, along with the brittleness, ossification, and systemic dependencies go away, it just hides them. :-) Russ _______________________________________________ sidr mailing list sidr@ietf.org https://www.ietf.org/mailman/listinfo/sidr