[OPSAWG] Review of draft-zheng-opsawg-tacacs-yang-01

[Joe Clarke <[e-mail]>]

As promised at the mic during opsawg at IETF 104, here is my more
detailed review of this draft.

As I stated during the meeting, I think the AAA module should be taken
out of this document.  I believe Alan has commented the same.  A AAA
module may be required, but I don't want to muddle the TACACS+ work with
that.  Plus, I'm not convinced opsawg would be the correct place for a
more general AAA module.

Secondly, I like the fact that you're extending the ietf-system module
in a manner similar to RADIUS.  I think that this work fits nicely there
for device admin.  I would remove the AAA moniker from the module for
now.  Leave it as ietf-tacacs-plus.

I recall seeing a comment on-list that any reference to "tacacs" must be
"tacacs+" or "tacacs_plus" or similar.  TACACS without the plus is a
very different beast.  Let's not confuse what we're trying to do here.

Maybe I'm being overly pedantic here, but why is "options" separated
from other rw objects by the statistics branch?  I would think you'd
want to group the rw objects together.

The word "accounting" is misspelled throughout this document.  In
general, I would run a spell checker over it.  After listening to
Heather at the keynote, we should do our best to help out RFC Editor,
even early on in the document lifecycle.

What is the intent of network-instance?  Is this like specifying a VRF
on which to reach the T+ server?  The description was not very clear.

You have a source IP option, but I know some vendors also implement a
source-interface.  I think it would be useful to have that as well
(maybe a choice there).

Joe