RE: [802.1] RE: I-D ACTION:draft-ietf-bridge-8021x-02.txt

Hi folks,
   At the July Plenary we have decided to put the 802.1aa mib back to its
original 802.1X state (or as close to it as we can achieve).  Those values
which are EAP or AAA related (counts of 'request ids' and 'response ids',
etc) will have their text changed to indicate that these values will be
filled by a functional entity other than the 802.1X PAE(the EAP entity for
instance).
   The reason for this issue is that .1aa clearly dilineates labor between
the 802.1X PAE and EAP and AAA.  The guiding principle in .1aa is that it is
to transport EAP and switch between controlled/uncontrolled port, not
interpret EAP, this means it should not look at the EAP header, only the
EAPOL header.  With this principle, the PAE is unable to keep some MIB
statistics and those need to be handled by the appropriate entities
(although in actual implementations it is very likely that one piece of
software is handling PAE and EAP on the authenticator, the standards are
separate).
   Our options were to:
       1.  Alter the existing MIB to get the closest values to the current
ones while only looking at the EAPOL header.
       2.  Remove all EAP related values and let the EAP group define their
own MIB.
       3.  Leave the MIB alone and let the text indicate that some values
need to be filled by entities other than the PAE.
   We chose #3 because we don't want to change the MIB for fear that the AP
and Bridge manufacturers may not update their MIBs to the new one in a
timely fashion and then management software will need to deal with these
different MIBs leading to potential interoperability issues.  The feeling is
that it took a long time for the industry to get the current MIB implemented
properly on their equipment, this MIB is extremely important to
troubleshooting and we don't want to add yet another potential complexity to
wireless access.  So, the MIB remains the same, the definition for the
values remain the same, but which functional entity is expected to fill some
of the values is made more clear.
   The changes to bring the .1aa MIB back toward the original .1X MIB should
be done soon.
   If you would like to review the changes please let me know and we can
send them out to you as soon as we get them done.  These will be 'beta'
changes as they will still need to go through the IEEE balloting process,
but it would allow you to track it and give input.
Thanks,
Jim B.

-----Original Message-----
From: [email protected]
[mailto:[email protected]]On Behalf Of Les Bell
Sent: Thursday, July 31, 2003 4:08 AM
To: C. M. Heard
Cc: [email protected]; [email protected]
Subject: [802.1] RE: [Bridge-mib] I-D
ACTION:draft-ietf-bridge-8021x-02.txt






The changes to the MIB for IEEE 802.1aa include the corrections in
draft-ietf-bridge-8021x-02.txt.

There are also some other significant changes to be done for 802.1aa, due to
recent changes in the state machines and the managed objects that reflect
this.
This work is not complete yet

Les....





"C. M. Heard" <[email protected]>@ietf.org on 30/07/2003 20:30:07

Sent by:  [email protected]


To:   [email protected]
cc:   [email protected]
Subject:  RE: [Bridge-mib] I-D ACTION:draft-ietf-bridge-8021x-02.txt


On Wed, 30 Jul 2003, Wijnen, Bert (Bert) wrote:
> An alternative option might be to publish the RFC without the MIB
> module and just include a pointer to the inline MIB module at the
> IEEE web site.

Actually, I rather like that idea.  It certainly ensures that the
stuff in the information RFC can't get out-of-sync with the IEEE's
official version.  The only problem with that

http://www.ieee802.org/1/files/public/MIBs/802-1x-2001-mib.txt

has some problems that make it fail to compile (an illegal expression
for the OID assigned as the MODULE-IDENTITY value, non-ascii quotes,
and the absence of dot1xPaePortReauthenticate and
dot1xAuthSessionUserName from conformance groups).  The version in
the -01 draft corrected these problems.

I see that the IEEE is in fact working on a maintenance release of
802.1X (see http://www.ieee802.org/1/pages/802.1aa.html, or look for
the 802.1aa link under http://grouper.ieee.org/groups/802/1/), and
one of the work items is a MIB module update.  However, this is in
the private area and I don't know whether it includes the
corrections mentioned above (nor how else it might differ from the
current published version of the IEEE8021-PAE-MIB).

Maybe someone from 802.1aa could comment?

//cmh


_______________________________________________
Bridge-mib mailing list
[email protected]
https://www1.ietf.org/mailman/listinfo/bridge-mib