Re: suggestion for new ssh maintenance wg

On Fri, Jan 29, 2016 at 2:49 PM, denis bider <[email protected]> wrote:
>> SFTP v3 - as implemented by OpenSSH and everyone who wants to interoperate
>> with it
>
> Gah, sorry, wrong version. That's SFTP v4, which is also implemented by many
> clients and servers. However, most that implemented v4 by now also implement
> v6.
>
> This is version 3:
>
> https://tools.ietf.org/html/draft-ietf-secsh-filexfer-02
>
>
> denis bider <[email protected]> , 1/29/2016 9:59 PM:
>
> With regard to AEAD:
>
> I think we should just make the following simple and clear statement:
>
> MAC algorithms are secondary to encryption algorithms, and are evaluated
> only if the encryption algorithm is not AEAD. If an AEAD encryption
> algorithm is negotiated, the outcome of MAC negotiation is irrelevant and
> must be ignored. If no mutual MAC algorithms are available, this causes key
> exchange to fail if, and only if, the negotiated encryption algorithm is not
> AEAD.
>
> I believe this is what [email protected] does, and is the behavior that
> makes most sense to me.
>
>
> With regard to SFTP:
>
> We may not have an agreement in what SFTP should be; but by necessity, we
> have an agreement in what it is.
>
> I suggest we should simply do the minimum work possible to promote the
> following drafts to some type of RFC, and call it a day:
>
> SFTP v3 - as implemented by OpenSSH and everyone who wants to interoperate
> with it
> https://tools.ietf.org/html/draft-ietf-secsh-filexfer-04
>
> SFTP v6 - implemented by many (perhaps most?) clients and servers
> https://tools.ietf.org/html/draft-ietf-secsh-filexfer-13
> https://tools.ietf.org/html/draft-galb-filexfer-extensions-00
>
> This is the de facto SFTP as we have it. Implementers have to follow these
> drafts, so they might as well be RFCs.

Apprently OpenSSH isn't implementing v6, and after looking at it I see
why, although if some other people see uses for it, so be it. If these
aren't RFCs given how important and widely used sftp is (and there
isn't even a draft for the even more useful scp), it raises lots of
questions about what the IETF has been doing, and whether it's worth
it.

>
>
> ----- Original Message -----
> From: Niels "Möller"
> Sent: Friday, January 29, 2016 15:34
> To: Stephen Farrell
> Cc: [email protected] ; denis bider ; Watson Ladd ; Daniel Migault ;
> Curdle Chairs ; [email protected]
> Subject: Re: suggestion for new ssh maintenance wg
>
> Stephen Farrell <[email protected]> writes:
>
>> If you think such an ssh maintenance wg is a bad plan,
>> please also do say that and why you think that.
>
> There's definitely some work that needs to be done. I'm not very
> familiar with ietf processes, so I'm not sure a new working group would
> make it easier to make progress. I guess what's needed is either an
> active wg chair, or an active area director, or someone informally
> accepting (and being accepted) in a similar role.
>
>> PPS: Note that this could be short-lived wg that never
>> needs to meet face-to-face, or maybe it'd not be like that,
>> but don't get fussed about having to go to IETF meetings
>> to get this work done - if it's maintenance then that may
>> well not be needed.
>
> Don't worry about IETF meetings. I felt I was deeply involved during the
> work on the ssh rfc:s. And I've never been to a secsh wg meeting, only
> on the mailing list. (I've actually been to one ietf meeting in my life,
> but the secsh wg didn't meet that time).
>
>>> Extension negotiation for SSH:
>>> https://datatracker.ietf.org/doc/draft-ssh-ext-info
>
> An extension mechanism makes sense to me, but I find most of the
> proposed extensions questionable and/or hard to get right.
>
>>> In addition to the above, I very much agree that [email protected]
>>> needs standardization.
>
> I think the single issue that might motivate forming a new wg is how to
> properly negotiate the use of aead crypt in ssh. There should be no
> difference between aes-gcm (which I'm not very fond of) and
> chacha-poly1305.
>
>>> Among other things, the erstwhile SSH working group never finalized
>>> the SFTP spec due to lack of consensus. We now have two SFTP specs,
>>> version 3 implemented by OpenSSH, and version 6 implemented by most
>>> everyone else.
>
> I honestly doubt we'll see much progress there, wg or not. It was a bit
> too much of second system syndrome. But if some others have the energy
> to revive it, I can't object, of course.
>
> Regards,
> /Niels
>
> --
> Niels Möller. PGP-encrypted email is preferred. Keyid C0B98E26.
> Internet email is subject to wholesale government surveillance.



-- 
"Man is born free, but everywhere he is in chains".
--Rousseau.