At Fri, 19 Sep 2008 16:09:46 +1000, Geoff Huston wrote: > > As usual, though, this does raise another question: If, at all times, > the EE cert is only used to sign the "current" manifest, then does the > EE cert require its own SIA and publication area?
No. > My inclination is to say that any point in time it looks like a > single use EE cert, its just that you MAY keep the private key on > hand, and MAY reuse the cert to sign a sequence of manifests.. > > Is this consistent with your view? Except for my continued confusion about the overloaded term "single use", yes. :) > > Section 2.1: First paragraph is confusing and flat wrong in a few > > places. I think this stems from attempting to talk about a "CA's" > > publication point rather than a "CA certificate's" publication point. > > Second paragraph has minor issues along the same lines. Suggested > > rewrite of the entire section: > > > > In the case of a CA certificate's manifest of its associated > > publication repository in the scope of the Resource Certificate PKI > > (RPKI), the manifest contains the current published certificates > > issued by this CA certificate, the most recent CRL associated with > > the CA certificate, and all objects that are signed using a > > "single-use" EE certificate issued by this CA certificate. > > > > In the case where an EE certificate has a defined publication > > repository, the EE certificate's manifest contains all published > > objects that have been signed by the EE certificate's key pair. > > > > Now i'm in a terminology soup too. > > I thought that it was appropriate to talk about a CA, where CA stands > for "Certification Authority", which is an entity that issues digital > certificates for use by other parties, so it was appropriate to talk > about a CA's manifest, CRLs issued by a CA and certificates issued by > a CA, > > You are proposing to use the term "CA certificate" instead. > > Now for me the "CA certificate" is not the active entity, its the "CA". The problem is that the manifest is a not a listing of the active entity's outputs, it's a listing of the outputs issued by a particular CA certificate. Now that David has finished beating us up about the term "rollover", the mapping from CA to CA certificate is 1:1, just to further confuse matters, and we lack a good term for the entity that controls a publication point and maps 1:1 with the publication point. There's a subtlety here that may not be obvious (thanks to Sandy for asking about it): the publication point, in the sense of a filesystem directory, is common to all current CA certificates in use by this nameless entity that we used to call a CA. Normally that's exactly one CA certificate, except during rollover, when both the old and new CA certificates are valid; now that David has whacked us with proper PKI terminology, the process that we call rollover is really a process of generating a new CA and reissuing all of the old CA's products under the new CA. The manifest, however, is specific to a particular CA certificate. This means that during rollover the manifest will not cover the entire content of the directory; instead, there will be two certificates, each with its own manifest, and the two manifests between them will cover the content of the directory. As the unnamed entity reissues subordinate objects under the new CA certificate, subordinate objects will move from the old CA certificate's mannifest to the new CA certificate's manifest, until the only thing left in the old CA certificate's manifest is the old CA certificate's CRL. In the fullness of time, the old CA certificate, its manifest, and its CRL all go away, the rollover event completes, and we're back to a single manifest that covers the entire directory. This model in which the directory is shared but the manifest is not may seem complicated, but it turns out to be straightforward to implement, and is necessary to keep the URIs in the SIA and AIA extensions stable in a hierarchical repository layout So the mapping from publication point to manifest is not 1:1, but everything in a publication point is covered by a manifest. In post-David terminology, the publication point is shared by the old CA and its successor (thus, necessarily by their respective CA certificates); the manifests are per-CA (thus, necessarily by their respective CA certificates). This should be the only case of shared publication points: an old CA and its successor(s), all of which are under the control of the unnamed entity that we used to call a CA. It's really the same thing we've been doing all along, with the terminology tweaked. I dunno if this puts us on the same page, but perhaps it helps to clarify the source of our confusion. Part of the reason why I prefer to phrase things in terms of "CA certificates" instead of "CAs" wherever possible is that a "CA certificate" is a small concrete object, while a "CA" is a complex entity with many moving parts and somewhat fuzzy around the edges. > > Section 2.2.1.4: This could probably be relaxed slightly. The MUST > > requirement is there to ease validation of manifests for CA > > certificates; it's not really necessary for manifest of EE > > certificates, as the validator already has the EE certificate in hand > > by the time it gets to the manifest in that case. The requirement is > > harmless in any case, so I'm ok with leaving this as it stands. > > this was a topic of a discussion some time back and the MUST was a > result of that conversation. Its not necessarily the case that the > validator has the EE certificate in hand - the CA MAY elect not to > publish the EE certificate separately and publish it ONLY in the form > of including it in the CMS wrapper of the manifest. So I thought that > the outcome was that the EE cert MUSt be in the CMS wrapper of the > manifest and MAY be separately published in a repository publication > point. I completely agree with the above for the case of a CA certificate's manifest, but this section was talking about EE certificates that sign multiple objects, including manifests. I'm having a hard time picturing not having the EE certificate in hand in such a case, but then we don't currently have anything that uses this particular capability so it hasn't yet passed the running code test. > So I propose to leave this as it stands, which I assume from your > comment above is still ok with you. Yep. > > Step 2 bullet list: > > > > First bullet requires an id-ad-signedobject SIA extension in the EE > > certificate. This is harmless, and could conceivably be useful, but > > is not really required given that the EE certificate will be bundled > > into the CMS anyway. I'm ok with leaving this as it stands, am just > > pointing it out in case it's an oversight from a previous edit. > > this is not so straightforward, as it now impacts on the res-cert > draft, section 3.9.7 > > Assuming that a sequence of manifests issued by a CA (and a CA has a > single key value) has a single publication point object name then the > id-as-signedobject SIA points to this object and all works fine. But > this assumption is a stipulation in the third bukket point of section > 3.3, so I guess that this is a long winded way of saying that even > with an EE certificate being used to sign multiple manifests in > sequence the id-ad-signedobject text is stillvalid and I propose to > leave that part of the text as it stands. > > Again, I assume from your comment above that this is ok with you. Yep. > > You will no doubt have noted that I'm using the term "one-off" instead > > of "single-use" here. This is not an accident. "Single-use" has > > proven to be ambiguous, in that it could mean a certificate that's > > used to sign multiple generations of a single object or a certificate > > that's used to sign exactly one generation of that same object. A > > better term than either might be "throw-away certificate", but that > > might be perceived as pejorative (which it's not, the technique is > > clever when used appropriately). > > again I agree that this concept, that a certificate is in use once at > any time, but is used multiple times in series is indeed ambigiuous > with the term "single use" > > I too would be loathe to use "throwaway" certificate, and I would liek > to keep the term "single use" in a fashion that is consistent with res- > certs, but I will think about this some more and see if there is some > text at the start of the manifest draft that could draw the > distinction between "single-use, used once only", and "single-use, but > could be reused multiple times" EE certificates If you think you can salvage the term, go for it. I doubt it can be done, since the term has two different meanings, each of which is self-evident to the reader, depending on context and the reader's preconceptions. That can be fun when writing poetry but is usually a bad sign in technical writing. > > Fifth bullet item: s/the old CA associated/the old CA certificate > > associated/ (the distinction matters, particularly in this paragraph). > > well I'm still having trouble with this one, terminology-wise! Substance covered above, I hope. Terminology usage details will require another review pass now that David has made us change terminology (not a complaint, just an observation). > > Section 4: There are several occurance of a rather odd phrase: > > > > The issuer SHOULD honour these values in the issued certificate or > > MUST reject the Certificate Request. > > > > I don't know what this means. Either it's a SHOULD or it's a MUST, > > can't have it both ways. If there's something more subtle going on > > here, somebody's going to have to explain it to me, preferably by > > rephrasing all of the occurances of this incantation, because it's > > totally opaque to me. > > ok - the ISSUER MUST honor the values in the issued certificate says > that the issuer has no choice but to honor them. = i.e. the issuer > cannot reject the request if the MUST truly is a MUST. The proposed > wording says - either issue the certificate as requested (SHOULD > honor) OR reject (MUSt reject). No other option is possible > > So I would say that SHOULD honour or MUST reject is appropriate. But I > would be interested to hear your suggestion of alternate wording that > conveys the same intent more clearly. Ok, I think I see what you're trying to say. Try: The issuer MUST either honor these values in the issued certificate or reject the request entirely. If that's too terse, try: The issuer MUST either honor these values in the issued certificate or reject the request entirely; the issuer MUST NOT issue a certificate that does not honor these values. I prefer the first of these, but either would be ok. I don't think the SHOULD is really necessary at all, since it's really just saying that the issuer SHOULD issue certificates when it can. > > Section 6.6: I find the discussion of the interaction between manifest > > and CRL here a bit disturbing for the case where the manifest is stale > > and the CRL is current. Everything this section says is correct, but > > one could easily come away with the belief that it's better to believe > > an old stale copy of the CRL than to believe the current CRL, just > > because the (stale) manifest matches the (old, stale) CRL. I don't > > think we want to prefer stale CRLs over current ones when we have both > > and both are valid, so there's some clarification needed here. > > Does the "this is a matter of local policy" apply in such cases? Sounds right. It's a tradeoff between currency and consistancy. _______________________________________________ sidr mailing list [email protected] https://www.ietf.org/mailman/listinfo/sidr
