---------- Forwarded message ----------
From:
Geoff Thompson <gthompso@xxxxxxxxxx>Date:
Fri, Sep 12, 2008 at 1:27 PM
Subject: 8023-POEP D3.1 additional comment
material
To: "D. Matthew Landry" <
dmlandry@xxxxxxxx>,
wael.diab@xxxxxxxxxCc: Geoff Thompson
<
gthompso@xxxxxxxxxx>,
david_law@xxxxxxxxColleagues
-
I have finally done some text generation to provide meat for the PoE+
additions to
Annex 30A and 30B. I did them in Microsoft word. I tried Frame
but it has been too long since I have done style controlled and heavily
reference linked editing in Frame and it was too big a hill for me to climb in
the time I had available.
My contribution is attached as 3
files
File named: 8023-30a_b_c.doc
- Contains the additions to the oPSE object in what I believe is the
appropriate place.
- The new text is shown in red underscore.
- The easiest way to find it is to open the file and then search on "oPSE".
That will put you in the immediate neighborhood.
File
named: oPD Cl30A-scratch.doc
- Contains the additions for the new oPD object to be added to the main file
in the appropriate place.
File
named: PoEPlus Cl30B-scratch.doc
- Contains the additions for the new ENUMERATIONS required to support PoE+
to be added to the main file in the appropriate place.
In the
process of actually reading more of this stuff than has been customary with the
rest of the world, I noticed some things that I think need fixing. If I had done
all ofthis when a ballot was open I would have actually entered comments into
the database.
Specifically:
- We have attributes that occur in both the PSE end and the PD end. These
have not been given unique names. It is my understanding that this will not
work and that when they end up in a MIB somewhere there would be no way to
tell them apart.
- - aLostCommunication
- - aMirroredLostCommunication
- are the ones that I noticed in particular.
- It looks like the "APPROPRIATE SYNTAX" ResourceInfo in 30.9.2.1.14 needs
work
- That is, the term "ResourceInfo" is (I think) not actually defined. We
need to define it or perhaps migrate to "ResourceTypeID" which is a
definition used in Repeater Management. The definition itself is one that is
imported from 802.1F if memory serves me correctly.
- If we define it ourselves then the definition should be more crisp (and
defined in 30B), e.g. a sequence of other syntaxes like
- Foo::= SEQUENCE {
- OUI [1]
- Which we would have to define or import
- String [2]
- }
- I think allowing an arbitrary number of strings is problematical.
See you in Seoul (or at the airport)
Geoff