Elwyn,

Nits/editorial comments:
s4, s4.1 and s4.2:
The  term 're-issue' is used in this section regarding the new
certificates that the CA creates, e.g. in s4.1:
When a CA, as part of the key rollover process, re-issues a CA
   certificate,
The term 're-issues' might be slightly confusing. Its a new certificate
but with lots of fields the same.  I would be inclined to go with
'issues new' but its a quibble.

There is not uniform terminology for describing this sort of process in the PKI world, but the term "reissue" (or re-issue) is reasonable. (It is distinct from cert "renewal," which implies a new cert with the same key and all attributes remaining constant except for the validity interval and serial number. Cert "re-key" refers to issuing a new cert with a new key, new serial number, maybe a new validity interval, but the same Subject name and other attributes remaining constant.) Also, we use the same terminology in several other SIDR docs, e.g., draft-ietf-sidr-repos-struct-07.txt, draft-ietf-sidr-ta-07, draft-ietf-sidr-res-certs-18, and in an analogous context in draft-ietf-sidr-rpki-manifests-11.txt.

Steve

---------------------

Bert,

Responses to your comments below:

...

RFC5706 suggests but does not require a separate
"operational considerations section".

The draft-ietf-sidr-keyroll-06 discusses the operational considerations
in the "Security Considerations" section. I can live with that and leave it
to the OPS ADs if they rather see that split out in a separate section
or not.

Thanks.


Substantive comments/questions:

- last para on page 4

   During the transition of the "current" and "new" CA instances it is
   necessary for the "new" CA instance to re-issue all subordinate
   products of the "current" CA.  The procedure described here requires

   I wonder if the "requires" should be a "REQUIRES" (yep, I know that
   rfc2119 does not have that exact capitalized form of require, but I
   think that a MUST or REQUIRED is intended here, no?)?

How about:

    During the transition of the "current" and "new" CA instances the "new"
CA instance MUST re-issue all subordinate products of the "current" CA.

- page 6
      4.  The NEW CA enters a Staging Period.  This duration of the
          Staging Period is determined by the CA, but it SHOULD be no
          less than 24 hours.
   Does/would it make sense to also suggest/recommend a max period?
   Would 48 hours make sense?

A recommended max period is probably a good idea, since relying parties are
burdened by having to deal with two sets of signed objects. If my co-author does not object, I would revise the text as follows:

4.        The NEW CA enters a Staging Period.  This duration of the
          Staging Period is determined by the CA, but it SHOULD be no
          less than 24 hours.  (It is RECOMMENDED that the staging period
          be no longer than 48 hours.) ...



- given the first para in sect 1, I wonder if ID.ietf-sidr-arch should be
   a normative instead of a informative reference.

OK, that's an easy change.

Editorial/nits:

- sect 1. 1st sentence
   This document describes an algorithm to by employed by a

   s/by/be/  (the first occurrence of "by")

OK.

- page 4, at top:
  CURRENT:
      The CURRENT CA is the active CA instance used to accept and
      process certificate issuance and revocation requests The starting

  s/requests/requests./

OK.

- page 5
   expand acronym CPS on its first use in this document.

     used in the CURRENT CA certificate.  The Certificate Practice
     Statement  (CPS) for the issuer of the NEW CA certificate will
     indicate the time frame within which a certificate request is
     expected to be processed.



- same for AIA acronym on page 6.

      -  Generate a new key pair, and generate an associated EE
         certificate request with an Authority Information Access (AIA)
value of the NEW CA's repository publication point.

- page 6:
  4.  The NEW CA enters a Staging Period.  This duration of the

  s/This duration of the/The duration of this/

OK

- page 8
   expand acronym CMS on its first use in this document.

     (If the signing time or binary signing time values in the Cryptograpic
     Message Syntax (CMS) wrapper are non-null, they MAY be updated to
     reflect the current time.)


Steve
_______________________________________________
sidr mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/sidr

Reply via email to