Re: [OPSAWG] I-D Action: draft-ietf-opsawg-tlstm-update-03.txt

[Kenneth Vaughn <[e-mail]>]

I just have a couple of comments on the proposed text:
1. Is it appropriate to mark 0-2 as reserved when they are technically assigned for deployments using TLS 1.2? Given that this is an update to the existing RFC, it seems as if the new design should define the values even if they should never be used. I propose that we either 
   a. provide a description along the lines of "obsolete - historically md5" or
   b. provide a description as defined in the original table (e.g., "md5") and adding a "Recommended" column with a value of "N" for entries 0-2. 

The second approach seems to be used by at least the TLS WG, so I am inclined to go with that unless I hear otherwise.

2. Perhaps I am missing something, but I do not think that the DTLS-OK column has any practical meaning within the scope of our document. I propose omitting this column as it could create confusion and inconsistencies.

Regards,
Ken Vaughn

Trevilon LLC
6606 FM 1488 RD #148-503
Magnolia, TX 77354
[phone]
[phone] cell
[e-mail]
www.trevilon.com

> On May 9, 2022, at 4:16 AM, tom petch <[e-mail]> wrote:
> 
> From: tom petch <[e-mail]>
> Sent: 06 May 2022 13:05
> 
> From: OPSAWG <[e-mail]> on behalf of Kenneth Vaughn <[e-mail]>
> Sent: 05 May 2022 19:19
> 
> <tp>
> following up my previous e-mil, I would make the IANA Considerations something like
> 
> =============================
> IANA is asked to created a new registry 
> SNMP Hash Algorithms
> in the 
> Structure of Management Information (SMI) Numbers (MIB Module Registrations)
> Group.
> 
> The registry has fields value, description, DTLS-OK and reference.
> 
> The range of values is zero to 255.  Values zero to two are reserved as are values 7 to 223.  Values 224 to 255 are reserved for private use.
> 
> The initial entries for values three to six are listed in section 2.1 of this document.
> 
> 
> The policy for updates is Expert Review.  The expert should consult the Security Area, e.g. via the mailing list of the TLS WG (the initial values of this registry are taken from an existing TLS Registry so the TLS WG would seem the best fit for this).
> 
> In addition, a new entry MUST be added to the SNMP-TLSTM HashAlgorithm Registry every time a new hash algorithm is approved for any version of TLS or DTLS.  A separate entry MUST NOT be created when an existing hash algorithm is used as a part of a new (D)TLS cipher suite.
> ===============================================
> with section 2.1 including the values from the current regisry.
> 
> Tom Petch
> 
> 
> 
> I have no real problem either way; unless I hear someone else argue for making it automatic in the next couple of days, I will delete the last two sentences of the first paragraph of the IANA Considerations so that it will read:
> 
> This document requires the establishment of a new SNMP-TLSTM HashAlgorithm Registry, which is referenced in the above MIB as being located at "http://www.iana.org/assignments/tlstm-parameters/". The initial values for this table MUST be identical to the contents of the TLS HashAlgorithm Registry (RFC 5246).
> 
> While future additions to the IANA TLS HashAlgorithm Registry are not expected, any future addition to the IANA TLS HashAlgorithm Registry MUST be consistent with the values assigned in the proposed IANA SNMP-TLSTM HashAlgorithm Registry.
> 
> <tp>
> Problem one;  the IANA website is a mess because IANA do what they are told:-(  This makes data difficult or impossible to find.  The RFC on IANA now calls for a two part structure of group and registry.  Here a group exists for SMI and that is where anyone should expect to find this registry so the instructions should then ask for the creation of a algorithm registry in the
> Structure of Management Information (SMI) Numbers (MIB Module Registrations)
> Group (alongside such as
> SMI Security for Cryptographic Algorithms).
> You have already seen how familiar security experts are with the abbreviation TLSTM.
> 
> Problem two; the registry you suggest for initial values includes MD5.  If that gets past IESG then it is time to abandon any pretence of security.  I think that only four values are currently appropriate.
> 
> Problem three; a registry needs an update policy.  Expert Review might be suitable.  It is up to IANA and the IESG to find experts, not us.
> 
> Tom Petch
> 
> 
> Regards,
> Ken Vaughn
> 
> Trevilon LLC
> 6606 FM 1488 RD #148-503
> Magnolia, TX 77354
> [phone]
> [phone] cell
> [e-mail]<mailto:[e-mail]>
> www.trevilon.com<http://www.trevilon.com>
> 
> On May 5, 2022, at 11:31 AM, Jürgen Schönwälder <[e-mail]<mailto:[e-mail]>> wrote:
> 
> We made the mistake by simply reusing the TLS hash algorithm for a
> different purpose. We now factor things out by having a separate
> registry for the hashing algorithms used to create certficate
> fingerprints. But why would we now tie this back to TLS hashing
> algorithms? In modern TLS, I think they only register entire
> ciphersuites, do we really expect them to go an register any new hash
> algorithm that comes along into the registry of hash algorithms for
> producing certificate fingerprints?
> 
> I understand the "we are too lazy to do this" argument but then still
> I would expect that if a new hash is being implemented for certificate
> fingerprinting, someone would find the energy to register it.
> 
> /js
> 
> On Thu, May 05, 2022 at 11:15:46AM -0500, Kenneth Vaughn wrote:
> There is no definition of TLSTMv1.3 nor do we version MIB modules
> Agreed, This is old text that I missed from when this was intended to be a replacement to RFC 6353 rather than an update. I think it is best to just delete the sentence so the paragraph would now read "[RFC6353] stated that TLSTM clients and servers MUST NOT request, offer, or use SSL 2.0. [RFC8996] prohibits the use of (D)TLS versions prior to version 1.2." While the statement is not technically required as it is stated elsewhere, I believe there was a comment during IETF 113 that we should be explicit about this.
> 
> In addition, a new entry
> MUST be added to the SNMP-TLSTM HashAlgorithm Registry every time a
> new hash algorithm is approved for any version of TLS or DTLS.
> I guess there are two issues here: 1) What do we want and 2) What wording is used for IANA instructions
> 
> In the first case, while I accept that there is not a strict technical requirement to implement every hash algorithm adopted by TLS, I am hard pressed to think of why we would ever not want to support one (or what practical harm it would cause). If we don't make the cross-assignment automatic, it seems as if it would be incumbent upon the WG to explicitly make requests every time a new hash algorithm came along. That seems to be extra bureaucratic work for OPSAWG and it seems as if it would be easier to make it a part of the IANA process. So the proposal is that it should be automatic (which is in agreement with a comment made at the IETF 113 meeting)
> 
> If we agree that it should be automatic, then the second issue is how do we state this. I am happy to revise the wording as appropriate; I certainly am not an expert in writing those statements.
> 
> If we don't want it to be automatic, then perhaps we need to reach consensus on how new entries will be added.
> 
> Regards,
> Ken Vaughn
> 
> Trevilon LLC
> 6606 FM 1488 RD #148-503
> Magnolia, TX 77354
> [phone]
> [phone] cell
> [e-mail]<mailto:[e-mail]>
> www.trevilon.com
> 
> On May 5, 2022, at 10:32 AM, Jürgen Schönwälder <[e-mail]> wrote:
> 
> Before I go and check the details...
> 
> [...] TLSTMv1.3 MUST only be used with
> (D)TLS version 1.2 and later.
> 
> What does this MUST tell me? There is no definition of TLSTMv1.3 nor
> do we version MIB modules. I understand the intention of the statement
> but we need to be more careful about the wording.
> 
> And what about this:
> 
> [...] In addition, a new entry
> MUST be added to the SNMP-TLSTM HashAlgorithm Registry every time a
> new hash algorithm is approved for any version of TLS or DTLS.
> 
> Why would that be a MUST? The SnmpTLSFingerprint is used by the MIB
> module to hash certificates and as such this hashing has nothing to do
> with any TLS internal use of hash algorithms. The reuse of the TLS
> hash registry back then was a matter of convenience, not a matter of
> having a strong binding to the TLS internal usage of hash algorithms.
> 
> /js
> 
> On Thu, May 05, 2022 at 10:09:45AM -0500, Kenneth Vaughn wrote:
> I have uploaded a new version of the "Updates to the TLS Transport Model for SNMP". This version includes the following changes:
> Changed the name of the registry to the SNMP-TLSTM registry
> Updated reference to DTLS 1.3 to reflect the publication of RFC 9147
> Clarified the first paragraph of Conventions to indicate that references to TLS, DTLS, (D)TLS, and TLSTM are version neutral except where specific versions need to be cited.
> Changed "SNMPv3" to "SNMP" in several locations where the specific version reference was unnecessary with our convention statement
> Indicated that Additional Rules for TLS 1.3 "may additionally apply to future versions of TLS"
> The document has been through several review cycles and has also been vetted by the TLS WG. At this point, changes are primarily editorial and I believe it is stable enough to proceed to the next step of the approval process.
> 
> Regards,
> Ken Vaughn
> 
> Trevilon LLC
> 6606 FM 1488 RD #148-503
> Magnolia, TX 77354
> [phone]
> [phone] cell
> [e-mail]
> www.trevilon.com
> 
> On May 5, 2022, at 10:07 AM, [e-mail] wrote:
> 
> 
> A New Internet-Draft is available from the on-line Internet-Drafts directories.
> This draft is a work item of the Operations and Management Area Working Group WG of the IETF.
> 
>     Title           : Updates to the TLS Transport Model for SNMP
>     Author          : Kenneth Vaughn
> Filename        : draft-ietf-opsawg-tlstm-update-03.txt
> Pages           : 30
> Date            : 2022-05-05
> 
> Abstract:
> This document updates the TLS Transport Model (TLSTM), as defined in
> RFC 6353, to reflect changes necessary to support Transport Layer
> Security Version 1.3 (TLS 1.3) and Datagram Transport Layer Security
> Version 1.3 (DTLS 1.3), which are jointly known as "(D)TLS 1.3".
> This document is compatible with (D)TLS 1.2 and is intended to be
> compatible with future versions of SNMP and (D)TLS.
> 
> This document updates the SNMP-TLS-TM-MIB as defined in RFC 6353.
> 
> 
> The IETF datatracker status page for this draft is:
> https://datatracker.ietf.org/doc/draft-ietf-opsawg-tlstm-update/
> 
> There is also an HTML version available at:
> https://www.ietf.org/archive/id/draft-ietf-opsawg-tlstm-update-03.html
> 
> A diff from the previous version is available at:
> https://www.ietf.org/rfcdiff?url2=draft-ietf-opsawg-tlstm-update-03
> 
> 
> Internet-Drafts are also available by rsync at rsync.ietf.org::internet-drafts
> 
> 
> _______________________________________________
> OPSAWG mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/opsawg
> 
> 
> 
> _______________________________________________
> OPSAWG mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/opsawg
> 
> 
> --
> Jürgen Schönwälder              Jacobs University Bremen gGmbH
> Phone: [phone]         Campus Ring 1 | 28759 Bremen | Germany
> Fax:   [phone]         <https://www.jacobs-university.de/>
> 
> 
> 
> --
> Jürgen Schönwälder              Jacobs University Bremen gGmbH
> Phone: [phone]         Campus Ring 1 | 28759 Bremen | Germany
> Fax:   [phone]         <https://www.jacobs-university.de/>
> 
> 
>