Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

PASS: *** SEC Motion to forward 1000BASE-T(IEEE 802.3ab) to REVCOM

Howard Frazier - Disapprove
Bob Grow - Approve
Paul Nikolich - Approve
Buzz Rigsbee
Bill Lidinsky - Approve
Dave Carlson - Approve
Geoff Thompson - Approve
Bob Love - Approve
Jim Mollenauer - Approve
Chip Benson - Approve
D Vaman - Approve
Ken Alonge - Abstain
Vic Hayes - Approve
Pat Thaler - Disapprove
Robert Russell - Approve
Bob Heile - Approve
Roger Marks - Abstain

Approve:  12
Do Not Approve:  2
Abstain: 2
Did not vote: 1

Our past experience with standards that were forwarded without any
implementations to verify their correctness has not been good.  This is a
lengthy and complex draft and the process of building and testing real
implementations is likely to uncover problems.  The responses to the ballot
comments on this just brush off the comments rather than resolving them.
While the document has come very far and I appreciate the efforts of the task
force, I cannot vote to forward it at this time.

Further, I strongly object to the categorization of my ballot comments as a
request to see "proof of concept."  That is not the issue I raised.  I believe
that the "concept" at least in broad strokes has been adequately proven.  My
concern is with the accuracy and completeness of the specification.
Implementation experience is the tool that lets us test whether, at a detail
level, the correct specifications have been included to allow interoperable
devices to be designed.  In implementation and testing, we find the specs that
are too tight to be met, the omissions, the contradictions and the errors that
occur in a complex specification.

IEEE has a tool for forwarding a document that is good but not yet well proven
in its details; a trial use specification.  This status would correctly inform
the reader of the status of this document.  It would set up a process for
collection of comments for the next two years as we gain implementation
experience.  I could support forwarding the document as a
trial use standard. I vote against the motion.
Pat Thaler

My position on this motion is based in part on a technical concern,
and in part on a procedural concern.

The technical concern was summarized in the comment I sent along with
my vote on the motion.  I believe that a demonstration of technical
feasibility, in the form of a working implementation, is vital. I
have stated this position consistently and unambiguously for about a
year now.

The procedural concern is that we (802.3 and the SEC) allowed the
802.3ab PAR to be approved with a "waiver" of sorts on the 5 criteria,
specifically the 4th criterion, Technical Feasibility.

According to the 802.3 operating rules, the points which must be
addressed for the Technical Feasibility criterion are:

 4) Technical Feasibility
    Demonstrated feasibility; reports - working models.
    Proven technology, reasonable testing.
    Confidence in reliability.

When the 5 criteria for 802.3ab were approved,  these points were modified
to read:

 4) Technical Feasibility
    Demonstrated feasibility
    Proven technology
    Confidence in reliability

Note the ommissions of the words "reports - working models"
and "reasonable testing".

The argument in favor of this modification was that it was impractical to
expect anyone to committ the required resources to develop a prototype
implementation so far in advance of the standard.  Thus, we relied on
simulations and analysis to assess the technical feasibility. While I
believe that this was a reasonable approach at the PAR approval stage,
I had every expectation that implementations would be designed and built
by private interests in parallel with the standards development process,
and that information learned from the implementation experience would
be presented and considered by the 802.3ab Task Force.

As recently as last November, 802.3 affirmed this expectation by adopting
the following motion, in response to my negative Working Group ballot:

  That the response to Howard Frazier's disapprove comment be modified to:

  "While we appreciate your concern, we expect the existence proofs
   to be available by Sponsor Ballot. Given the simulation results
   and the design experience, it is appropriate to go forward to
   Sponsor Ballot with the existing draft."

   M: Ms. P. Thaler S: Mr. K. Daines
   Y: 42 N: 0 A: 2 Approved.

My procedural concern, simply stated, is that the 802.3ab project has
failed to meet expectations.
Howard Frazier

I abstain.
In my judgement, the arguments that have been expressed against this motion
represent sound justification for a "no" vote. However, I don't know the
other side of the story. Roger Marks


Issue Date:   6May1999          Closing Date: 14May1999
Moved By:     Geoff Thompson
Seconded By:  Bob Grow

Move: Authorize forwarding 802.3ab (1000BASE-T) to RevCom for approval at the
June1999 IEEE-SA Standards Board Meeting contingent on the approval of a letter
ballot motion from the voters of 802.3 scheduled to close May 10.