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

Re: [10GBT] My final wrap-up presentation

Brad can probably add a lot more but here's my response -
1) Yes, guidance from Brad as to how the larger 802.3 body will view it is that it needs to be "single" PHY
2) The selection will be done by voting at the task force and the voters include PHY vendors, cabling/connector vendors, system vendors and many others so it is not just PHY vendors
3) I would like to see the multiple PHY proposers come together and put a consensus proposal on the table. If this happens we will check if there is >75% support and if there is, we have the basis of developing Draft 1.0 even though all the details you list may not have been worked out.
4) In the absence of #3, what gets selected depends on how people vote.
5) One set of channel models has been provided via the reflector. A few others will be provided soon.
I think all of the factors you suggest in your bullet list are probably being weighed by the participants and will factor in how they vote though the importance they attach to each item in you list probably varies significantly from individual to individual.
I highly recommend that interested parties ask for the information relevant to them either via the reflector or via other means.
As for item #3 in the above list,  I think the six proposals currently on the table have many similarities among them and I would encourage the proposers to see if they can reach consensus to come up with a common proposal. To summarize what I had mentioned earlier in my wrap up presentation, five of the six proposals are based on PAM. Five of the six are based on LDPC codes. There are differences in baud rate but these are small and I can see this and a number of other issues being resolved over time and possibly even after Draft 1.0.
If your proposal is on the wrong side of the majority on either of these fronts, you either need to be working really hard to convince others to change their minds or the chances that you will get to 75% are really low. Please note that my statement is not based on any philosophical belief that the majority is always "right". My statement is a reflection of how the process in the IEEE works, if >75% of the task force believes something is better, it will be accepted.
cell (650) 704-7686
office (408) 653-2235

From: stds-802-3-10gbt@IEEE.ORG [mailto:stds-802-3-10gbt@IEEE.ORG] On Behalf Of Jeff Warren
Sent: Wednesday, June 09, 2004 10:38 AM
Subject: Re: [10GBT] My final wrap-up presentation

Sanjay / Brad,
Great wrap up presentation. It sounds like draft 1.0 coming out of the July 2004 plenary is very optimistic.  
Since the 10GBASE-T Objectives will be met with a single 10GBASE-T PHY do you agree that draft 1.0 should be based on a unified "single" PHY proposal? If yes, what is the agreed upon IEEE process to arrive at a single PHY proposal?
With a goal of producing draft 1.0 coming out of the July 2004 plenary meeting (that's 5-weeks away) it seems like going into the July plenary meeting that this "10GBASE-T PHY proposal down-selection process" (i.e. from 5 to 1 or 2) should be agreed upon fairly soon via this reflector.
Should this process factor in:
  • closure on requirements - e.g. autonegotiation, latency, baud rate, channel model, etc.
  • maturity of PHY proposal - how long has it been on the table?,
  • complexity of PHY proposal,
  • system provider preferences, not just PHY vendors (customer gets to vote),
  • time to market - is tied to complexity of proposal,
  • consideration given to the most complete PHY proposals, when will samples be available?,
  • validation of spreadsheet PHY proposal numbers against a "finalize" channel model, 
    • break the pencil used to tweak the channel model,
  • cut off new proposals, that one seems obvious,
  • etc.....       
Maybe the place to start with on this "agreed upon down-selection process" is a combination of ranking the above list and adding to it. In either case it seems obvious that going into the July meeting this process needs to be defined, or you run the risk of making a sub-optimum decision or worse yet no decision that results in a delayed draft 1.0 standard.
      - Jeff 
----- Original Message -----
Sent: Wednesday, June 09, 2004 12:36 PM
Subject: [10GBT] My final wrap-up presentation

Here's a copy of my final wrap-up presentation from the Long Beach meeting.
Sanjay Kasturia
Editor-in-chief, 802.3an
cell (650) 704-7686
office (408) 653-2235
fax (408) 844-8187
Teranetics Inc.
2953 Bunker Hill Lane, Suite 204
Santa Clara, CA 95054