Re: [sidr] GOST & SIDR

[Samuel Weiler <weiler+lists.sidr at watson.org>]

On Fri, 17 Apr 2009, Randy Bush wrote:

> > And is there any reason why algorithm agility won't just work?
>
> how do you capability negotiate with a cert?  i understand how to do it
> with a protocol peer, but not a static object.

I'm not suggesting negotiation.

I'm asking "will RPKI partcipants who want to use algorithms different 
from the norm and/or their parents be able to do so without any bad 
effects"?  Using strange algorithms may well mean that most relying 
parties can't verify the certificates, but that's to be expected.

> i suppose with enough complexity, ...  but this is one of those time i
> think there is a version number in the protocol.

So long as "version number" isn't a synonym for "flag day", there's 
nothing wrong with having one.  But we already have algorithm 
identifiers in the certs.  Are those enough?

-- Sam