Re: [Lake] Call for adoption for draft-selander-lake-edhoc - respond by June 22

[Stephen Farrell <[e-mail]>]

Hiya,

On 22/06/2020 01:38, Christopher Wood wrote:
> Top-posting, since I think clarifying will help answer the comments
> below.
> 
> As observed below, cTLS is underway elsewhere in the IETF. That does
> *not* mean this WG needs to pick an entirely separate AKE. On the
> contrary, this WG ought to evaluate -- per the charter -- whether
> cTLS meets the needs of LAKE. If it does, and if folks favor that
> AKE, then the primarily outcome of this WG seems to be in formulating
> requirements which led to that decision. (That would be a fine
> outcome as far as I'm concerned. The success of a WG is not
> predicated on how many documents it produces.)

Hmm. I think both edhoc and cTLS could, in their final
incarnation, meet the requirements and that that's fairly
clear. What am I missing? I don't get what work needs to
be done to establish that, and assume that WG participants
can do whatever evaluation is needed for themselves.

> In sum, I think this adoption call is premature and goes against what
> the charter lays out. Let's do our due diligence here. 

I'm still confused about the basis for the above TBH.

> I strongly
> believe Introducing yet another AKE into the standards pool is
> something we should not do without serious contemplation.

That, however, is quite clear and seems to me a valid
basis from which to argue against adopting edhoc. (Noting
that "valid" does not mean "winning.")

Cheers,
S.


> 
> Best, Chris
> 
> On Sun, Jun 21, 2020, at 5:29 PM, Stephen Farrell wrote:
>> 
>> Hiya,
>> 
>> I've a couple of questions below as I don't quite get the basis for
>> your conclusion. (I'm not trying to argue your conclusion but I
>> don't understand it and would like to.)
>> 
>> On 22/06/2020 00:51, Christopher Wood wrote:
>>> I do not support adoption.
>>> 
>>> The charter of this WG states:
>>> 
>>> draft-selander-ace-cose-ecdhe is a candidate starting point for
>>> the LAKE produced by the WG. Any work available from TLS or other
>>> WGs that satisfies the determined requirements will also be
>>> evaluated for suitability, but does not preclude the WG from
>>> freely selecting its preferred LAKE for OSCORE.
>>> 
>>> Unless I missed it, work from TLS was not seriously evaluated
>>> for suitability.
>> 
>> I don't understand what work you mean. cTLS has been adopted by the
>> TLS WG and will be developed there. I don't understand what useful
>> thing this WG could do about that.
>> 
>>> In fact, it seems we jumped right over it and landed on 
>>> draft-selander-ace-cose-ecdhe.
>> 
>> Yes. That is the only other serious proposal of which I'm aware.
>> 
>>> The cTLS authors demonstrated that this variant of the protocol
>>> can indeed meet the requirements set out in
>>> draft-ietf-lake-reqs-04, without compromising any of the benefits
>>> that the TLS ecosystem brings to the table. (Support for
>>> different server authentication modes, for example, is something
>>> TLS is well equipped to support.)
>> 
>> That's fine, but doesn't seem to speak to this adoption call.
>> 
>>> 
>>> I understand that the TLS WG adopted cTLS and will continue its 
>>> development there. However, that does not seem relevant for what
>>> this WG chooses.
>> 
>> I really don't understand that. We have 2 serious proposals. One
>> will be processed elsewhere. The question before us now is whether
>> to pursue the other one or not. It seems entirely relevant to me
>> that we are left with one or zero things to work on in this WG.
>> 
>>> By analogy, this would be similar to QUIC developing its own key
>>> exchange protocol since UDP is just slightly different from TCP.
>>> 
>> 
>> I don't find that analogy that useful tbh. QUIC is a transport area
>> WG chartered to develop a transport protocol. This WG was chartered
>> to do work on the topic of key exchange, so I don't think the
>> analogy holds.
>> 
>>> Clearly, that was not the path chosen, and I think it would be a 
>>> mistake to do that here without seriously considering cTLS.
>> 
>> Again - I've no clue what "seriously considering" might mean. cTLS
>> is being worked on. If we do not adopt edhoc then what would this
>> WG be doing? I don't think is makes much sense for a WG to exist
>> merely to "seriously consider" a work item in another WG:-)
>> 
>> Cheers, S.
>> 
>> 
>>> 
>>> Best, Chris
>>> 
>>> On Sun, Jun 21, 2020, at 1:13 PM, Jesus Sanchez-Gomez wrote:
>>>> Hello All,
>>>> 
>>>> I support the adoption of this document.
>>>> 
>>>> I've worked with the technology in research projects/papers 
>>>> yielding good results.
>>>> 
>>>> There are several use cases where this technology is a good 
>>>> solution for different research projects at the University of 
>>>> Murcia and Odin Solutions.
>>>> 
>>>> While I've practical experience with this technology working 
>>>> specifically in LoRaWAN, its design makes it a good fit for any
>>>>  constrained radio link/LPWAN.
>>>> 
>>>> Also, it has potential to be implemented beyond LPWANs, like
>>>> more generic IoT scenarios with large scalability.
>>>> 
>>>> Best Regards,
>>>> 
>>>> 
>>>> -- Jesús Sánchez Gómez Contratado predoctoral // Phd Student. 
>>>> Fundación Séneca. Comunidad Autónoma de la Región de Murcia +34
>>>> 868 88 96 74 [phone] [e-mail] Department
>>>> of Information and Communication Engineering Faculty of
>>>> Computer Science University of Murcia 30100 Murcia, Spain
>>>> 
>>>> -- Lake mailing list [e-mail] 
>>>> https://www.ietf.org/mailman/listinfo/lake
>>>> 
>>> 
>> 
>> Attachments: * 0x5AB2FAF17B172BEA.asc * signature.asc
>