On Feb 12, 2016, at 5:19 AM, Eliot Lear <[email protected]> wrote:
> Actually the work started long before THAT, even.  I think it dates to
> 1993 if memory serves. 

  Pretty much.  RFC 2989 and RFC 3127 are the outcomes (~y2000), not the start 
of the process.

> But speaking plainly, this isn't a science experiment.  It isn't
> something anyone is floating as a suggestion.  It's what's being done in
> the network today, and it is supported across multiple implementations. 
> It is a defacto standard.

  Like IPX, or any number of other non-IETF protocols.

  That doesn't mean they deserve RFC status, much less a standards track 
document.

>  Unless there's very good reason not to do
> so[*], let's just call it what it is.  And again, I would reiterate:
> let's also make clear what it's not.  A good applicability statement
> should help with that.
> 
> Eliot
> [*] A very good reason would be if it doesn't meet the requirements for
> specifications that the IETF generally laid out.

  The discussion here has been that TACACS+ is a AAA protocol.  If I believe 
that, I also have to believe that it's such a terrible AAA protocol, that it 
wasn't even considered in the discussion in RFC 3127.

  i.e. at the time the original draft was published, the major networking 
companies (including Cisco via their employees) and the ITEF consensus was that 
TACACS+ was not suitable for consideration as an AAA protocol.

  We're welcome to re-visit that decision.  Priorities change, and we learn new 
things.

  But the cognitive dissonance here is mind-boggling.  People say that TACACS+ 
is a AAA protocol, but the requirements on AAA protocols don't apply, because 
it's not a AAA protocol.  It's a device management protocol.  Even though the 
document doesn't describe device management, or even contain the terms "device 
management" or "device administration".

  By that logic, war is peace, freedom is slavery, ignorance is strength.


  I've asked repeatedly why the document should be granted standards track 
status.  The argument is largely "it's widely used".  Well, so what?  Many 
other protocols are *more* widely used, and aren't standards track.

  Again, while there are many widely used protools as information, or not even 
RFCs, while there are inconsistent arguments in favour of the protocol, while a 
self-described AAA protocol goes against IETF consensus, while a self-described 
AAA protocol is so inadequate that it wasn't even considered in RFC 3127, while 
there were procedural issues making it a WG document...

  Why, exactly, are is the WG full-steam ahead in making it a standards track 
document?

  It looks like the IETF ideal of "individuals" is failing here.  There are 
large forces behind the scenes pushing for standardization, and that is winning 
over petty things like individuals without deep pockets, and IETF process.

  Alan DeKok.

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

Reply via email to