On 12/02/2016 09:14, Alan DeKok wrote:
> On Feb 11, 2016, at 2:31 PM, Brian E Carpenter <[email protected]> 
> wrote:
>>
>> As far as I can tell the only point at issue here is two words in the
>> third header line of the draft:
>>
>> "Intended status: Standards Track"
> 
>   For me, that's just the tip of the iceberg.  There isn't much point in 
> talking about more minor issues when there are major issues still unresolved.
> 
>> There doesn't seem to be any dispute that TACACS+ is a widely deployed
>> protocol
> 
>    I agree.
> 
>> and that this is a competent description of it,
> 
>   Sort of.  It's a description of the TACACS+ base protocol as a generic AAA 
> protocol.  Which *on it's face* has 100% overlap with RADIUS.
> 
>  The defence of TACACS+ is entirely related to use-cases which aren't in the 
> document.
> 
>  Even if we publish this document as a standards track document, the contents 
> of the device management queries and responses can *only* be vendor specific. 
>  Each vendor has their own command-line syntax which, while similar, have 
> proprietary differences.  So the *use-case* of the protocol remains entirely 
> vendor-specific.
> 
>   So the base protocol would be documented, but the practical application 
> thereof would be 100% vendor proprietary.  That's an unusual use-case for an 
> IETF standard protocol.
> 
>> and as already
>> observed the WG chairs didn't exceed their authority in the adoption
>> process. Since the WG hasn't yet even been asked (by WGLC) whether the
>> document should be approved, I can't see what appealable decision has
>> been taken, since basically no decision has been taken.
> 
>   Does this mean that WG adoption is a meaningless step?

Yes, in terms of the IETF process rules for approving a document are concerned.
The notion of formal WG adoption is useful innovation, but it isn't actually
required by the rules.

>  That charter updates are meaningless steps?

Charter updates are formal step, but the OPSAWG charter is already written
to encompass work of this kind. As I hinted before, it's unfortunate that
the WG milestones haven't been updated recently, but that's true of many WGs.
We are collectively sloppy on that.

>   I'm a little surprised to see a public decision by a WG chair be described 
> as "no decision".  That's... legalese I didn't expect in the IETF.

It's no decision on the disposition of the document. Until the end of a WG last 
call,
the document is completely open for discussion and change by the WG.

>> Could we perhaps have a rational debate about whether the draft should be
>> on the standards track or not? It's a perfectly valid question to ask.
> 
>   I've been asking for reasons why it should be adopted. The answers are 
> largely "it's widely used".

Actually I think you've been asking for reasons why it should standardised, and 
there
are only two words in the draft header that affect that question. I don't 
really have
an opinion on that.

>   Well, so is IPX.  IPX has multiple implementations by multiple vendors, and 
> is running in probably 1000 times as many systems as TACACS+.  It's even 
> older than TACACS+.  It's still in use.
> 
>   Sure, IPX overlaps with IPv4, but heck... it's widely used, so let's make 
> it an IETF standard protocol.
> 
>   If you find those reasons unconvincing, you should find the reasons for 
> TACACS+ adoption unconvincing, too.

TACACS+ is used by IP operators, so it seems relevant in a way that IPX isn't.
But if you want to propose an IPX BOF in the Internet Area, or request a slot
in a future INTAREA WG meeting for IPX, you're obviously free to do so.

Rgds
   Brian

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

Reply via email to