David,
Even though your message did not include me in the salutation, it was
a reply to my message, so I'll take the liberty of replying.
I think the techniques you proposed for generating Subject names that
are unlikely to collide in a non-malicious context are good
suggestions. If the Wg agrees, we could add these suggestions to one
of the SIDR docs, as recommend (not mandated) name generation
techniques.
However, as I have explained previously, we cannot assume that there
will be no malicious name collisions, given the extent and the nature
of the RPKI. Thus we need to adopt defensive strategies, based on
crypto-validation of bindings, not just names. Because the RPKI
requires each CA to directly issue CRLs, its not hard to verify this
binding.
I agree that it is generally desirable to impose validation criteria
that is well-supported by extant tools like OpenSSL. However, I would
not make compatibility with that software dictate what we do. This is
especially true given that we have seen OpenSSL deviate from PKIX
standards in many ways over the years.
Steve
_______________________________________________
sidr mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/sidr
