Re: [sidr] Stephen Farrell's No Objection on draft-ietf-sidr-rpsl-sig-12: (with COMMENT)

[Stephen Kent <[e-mail]>]

Stephen, et al.,

A couple of observations about the topic of certs used to verify RPSL sigs:

     - the title of the I-D says that it relies upon the RPKI, and, as 
currently
written, it mandates use of RPKI certs. So, using certs from a different PKI
would require a re-write. Also, the security of the system would be reduced
if other certs were employed, wrt verification of assertions about address
space and ASN holdings. So, I don't think it's appropriate to suggest 
alternative
PKIs. I also note that the text on page 3 that says "equivalent 
functionality can
be achieved" using an alternative PKI is questionable. Merely having 
3379 extensions
in a cert does not mean that all of the other security-relevant elements 
of the RPKI
accrue. (Also, there is a typo: certificate authority" -> "certification 
authority")

     - if one is using RPKI EE certs, as currently mandated, then the 
single use
requirement (which is imposed on CAs, but RPs are not required to 
verify) applies.
Deliberate re-use of a cert would mean that the CA violated the cert 
policy, and
thus the cert ought not contain the RPKI policy ID, etc. Also, the 
Security Considerations
section of the I-D refers to 6487, which is appropriate only if we're 
discussing RPKI certs.

     - The phrase "money-in-the-middle" is cute, but Randy's terminology 
is not widely
adopted. I suggest using the more conventional terminology that a wider 
range of
readers will recognize.

Steve