RE: [EFM-P2MP] new opcode for DISCOVERY GATE - an open TR from last meeting
Gerry,
Besides speaking in metaphors about oil and water, we might look at the
practical perspective.
Most of our work in tweaking the state machines is either improving the
representation, or fixing bugs in the diagrams and clearly deals with the
"specification". This is not engineering of the protocol, or its options for
interaction - the "implementation".
The proposal for adding an Opcode is fixing a "specification" issue in Draft
1.3, which does not exist in Draft 1.414. You will see when reading Draft
1.414 that the proposed change does not in-fact simplify any diagram.
Besides having no benefit it has a terrible price, as it greatly impacts the
"implementation" - it actually breaks what was so far considered as a stable
protocol by all the vendors.
It is an injustice to describe this as 'small modification'.
It is also clear that affecting the "implementation" of the protocol while
not fixing anything that is broken is definitely a "new feature".
Ariel
Ariel Maislos
Passave Inc.
1557 Jasper Dr.
Sunnyvale, CA 94087
ariel.maislos@passave.com
Tel: (408) 530 0458 Cel: (408) 242 8599
-----Original Message-----
From: owner-stds-802-3-efm-p2mp@majordomo.ieee.org
[mailto:owner-stds-802-3-efm-p2mp@majordomo.ieee.org] On Behalf Of Gerry
Pesavento
Sent: Thursday, May 01, 2003 9:17 PM
To: ariel.maislos@passave.com; 'Glen Kramer'; stds-802-3-efm-p2mp@ieee.org
Cc: 'Howard Frazier (Howard Frazier)'; 'Wael William Diab'; 'Ben Brown';
'Yukihiro Fujimoto'
Ariel,
I believe your logic does not hold because we are talking about a GATE
process and a DISCOVERY process. They have different roles, and
different information to convey. This is clear from our most basic
functional block in 64-4, where we divide GATE, REPORT and DISCOVERY
functional blocks. A DISCOVERY GATE is mixing oil and water, and, for
example, forces us to list OLT parameters in the GATE message - which is
used to assign grants.
A clean protocol would have GATE and REPORT messages used for normal
operation, and nothing to do with discovery, registration,
de-registration, OLT parameter setting, LLID assignment, etc.
> We are beyond the deadline to introduce new features.
This is not a new feature. It simplifies our block and state diagrams.
Reviewing the Comments on D1.414, I see the state diagrams are still
being tweaked, so this in fact would be the time to do this. Like
Elvis says, it's now or never.
Gerry
> -----Original Message-----
> From: Ariel Maislos [mailto:ariel.maislos@passave.com]
> Sent: Thursday, May 01, 2003 7:43 PM
> To: 'Glen Kramer'; stds-802-3-efm-p2mp@ieee.org
> Cc: Gerry Pesavento; Howard Frazier (Howard Frazier); Wael William
Diab; Ben
> Brown; Yukihiro Fujimoto
> Subject: RE: [EFM-P2MP] new opcode for DISCOVERY GATE - an open TR
from last
> meeting
>
> Glen,
>
> The proposal you have made is attempting to fix a cosmetic problem in
Draft
> 1.3 that no longer exists in Draft 1.414.
> As a result the last couple of slides in your presentation are not
correct.
>
> What I do see in the proposal is a solution looking for a problem.
> For every flag we have defined in our protocol we can always define a
> separate message:
> 1) Why not make a separate message for Deregister?
> 2) Why not make a separate message for Force Report?
> 3) Why not make a separate message for Keep-Alive?
> 4) Why not make a separate message for Broadcast OLT Capabilities?
> This is an endless debate.
>
> Creating a new message for Discovery Gate adds no new functionality,
and
> fixes no bug in the draft.
> It is a change for the sake of change.
> At the stage the standard has reached I expect stability.
>
> We are beyond the deadline to introduce new features.
> Change for the sake of change is much worse.
> Especially when it is a large change as is introduced in this
proposal.
>
> Let's work together to complete the standard by converging on
stability, and
> not introducing changes at every opportunity.
>
>
> Ariel Maislos
> Editor, P2MP
>
> -----Original Message-----
> From: owner-stds-802-3-efm-p2mp@majordomo.ieee.org
> [mailto:owner-stds-802-3-efm-p2mp@majordomo.ieee.org] On Behalf Of
Glen
> Kramer
> Sent: Thursday, May 01, 2003 4:45 PM
> To: stds-802-3-efm-p2mp@ieee.org
>
> All,
>
> Currently, we have GATE opcode used for normal GATE messages
> as well as for DISCOVERY GATE messages. To distinguish the
> two, we have to use an additional Flag = {Normal|Discovery}.
>
>
> We have one open TR left from the last meeting that suggests
> making DISCOVERY_GATE a separate message from the normal
> GATE by adding a new opcode. Attached presentation explains
> the reasons and benefits of doing this.
>
> Please, let me know if you agree that this is a small but
> useful modification, and whether you support the idea.
>
> Thanks,
> Glen
>
>