FW: FW: FW: discussion about rogue database objects
Hi Rob, I've already explained why this had been happened at all first hand. RIRs have different rules and principles and constantly invents new methods of complicating the situation in exchanging information between them. Second thought: the "concern" was raised in presumption that the route objects "authorize" something to a provider. It is NOT. The absence of the routing objects would change nothing in the situation. A lot of providers do not build any filters based on a RIR routing policy at all. Then it just was not visible in the DB, but the routes were still here. And at this level no RIR with all its efforts cannot do anything (real routing decision is out of scope of any RIR). And this situation should go exactly in the law department of any company. Solving the problem with adding doubtful "security" is a waste of time and efforts. Regards, Vladislav -----Original Message----- From: Rob Evans [mailto:rhe@nosc.ja.net] Sent: Thursday, November 13, 2014 12:00 PM To: Potapov Vladislav Cc: routing-wg@ripe.net Subject: Re: [routing-wg] FW: FW: discussion about rogue database objects Hi Vladislav,
I'm against the methods of resolving "problems" which are not exist
I believe we've already heard of two instances of this problem existing -- the route objects that started this thread off, and Elvis Velea saying it has impacted at least one of his customers. Cheers, Rob
Den 13/11/14 10:09, skrev poty@iiat.ru:
Hi Rob,
I've already explained why this had been happened at all first hand. RIRs have different rules and principles and constantly invents new methods of complicating the situation in exchanging information between them. Second thought: the "concern" was raised in presumption that the route objects "authorize" something to a provider. It is NOT. The absence of the routing objects would change nothing in the situation. A lot of providers do not build any filters based on a RIR routing policy at all.
Yes. This was also the case with the providers of the rogue AS that started this discussion. There were even prefixes in the RIR db with a different origin AS than the one being announced on the internet. The data in the RIR db seem to have no influence on the routes being announced. You could however argue that a more correct RIR db would make LIRs use this db for actual filtering.
participants (2)
-
Jørgen Hovland
-
poty@iiat.ru