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

Re: [BP] Signaling spread sheet



I have the following proposals.

 

1) That we categorize channels into 3 categories.

à      Channels which the objectives require us to support

à      Channels which desireable to support

à      Channels which contain anomalies and are relevant to explore signaling margin and potential issues

 

2) That the performance be summarized by the following metrics:

à      Number of Required Channels Passing at specified BER

à      Number of Required Channels Failing at specified BER

à      Number of Desired Channels Passing at specified BER

à      Number of Desired Channels Failing at specified BER

à      Number of Anomalous Channels Passing at specified BER

à      Number of Anomalous Channels Failing at specified BER

 

3) Until we are able to categorize the channels into categories as above that we summarize the results by those channel's that meet the Channel Ad Hoc Mask sets.

à      Number of Channel Ad Hoc Compliant Channels Passing at specified BER

à      Number of Channel Ad Hoc Compliant Channels Failing at specified BER

à      Number of Channel Ad Hoc Non-Compliant Channels Passing at specified BER Number of Channel Ad Hoc Non-Compliant Channels Failing at specified BER         

 

 

 

In addition I believe that since time is limited, the signaling Ad Hoc should only pursue those channels which have met the 802.3ap schedule and have provided complete proposals.

 

There are two such complete proposals:

      Proposal for 10G Serial Backplane PHY Using Unified Signaling

            http://www.ieee802.org/3/ap/public/sep04/gaither_01_0904.pdf

            http://www.ieee802.org/3/ap/public/sep04/palkert_02_0904.pdf

 

 

      Proposal for 10Gb/s single-lane PHY using PAM-4 signaling

            http://www.ieee802.org/3/ap/public/jul04/brink_02_0704.pdf

 

 

I do not see a complete proposal for any other 10 gigabit serial solution.

 

 

 

 

Mike Lerer

Chairman Physical Link Layer Working Group of the Optical Internetworking Forum

Chairman Hardware Working Group Network Processing Forum

Box 636

Londonderry, NH 03053

Cell: 603-548-3704

-----Original Message-----
From: owner-stds-802-3-blade@listserv.ieee.org [mailto:owner-stds-802-3-blade@listserv.ieee.org] On Behalf Of Altmann, Michael W
Sent: Tuesday, November 02, 2004 11:37 AM
To: STDS-802-3-BLADE@listserv.ieee.org
Subject: Re: [BP] Signaling spread sheet

 

 Mike,

 

To start with, I think that I can state pretty confidently that here is

no implementation, coding, or channel bias in this spreadsheet ...

mostly because it's almost entirely blank!

 

I have used a worst-case assumption for summarizing the channel

performance, because I can only fit so much data on my screeen, and

nobody in the group has proposed another metric for summarizing this.

If a metric exists, which the group accepts, then I will happily enter

it into the spreadsheet.

 

I suggest that the best way to proceed, if this is not an appropiate

method of reducing channel BER results, is to propose another complete

and unambiguous metric for this comparison which the group can then vote

on.

 

.../Mike

 

-----Original Message-----

From: Mike-Lerer [mailto:mike@mike-lerer.com]

Sent: Tuesday, November 02, 2004 5:10 AM

To: Altmann, Michael W; STDS-802-3-BLADE@listserv.ieee.org

Subject: RE: [BP] Signaling spread sheet

 

In reviewing the spreadsheet, I believe that you are once again making

unreasonable assumptions about the interests of the majority of the task

group.

 

If I interpret the spreadsheet correctly, you are characterizing each

signaling scheme by its worst performance over any channel.

 

I believe that it is unreasonable to characterize a signaling scheme in

this

way unless or until the group makes a formal decision that each of the

channels falls within its required objectives.

 

As you will recall, at the last face to face meeting several motions

were

attempted to select channels for use in signaling evaluation, all of the

motions failed.

 

Until the group is able to make a definitive statement about which

channels

are required to be supported, desired to be supported, and merely of

interest as anomalous experiments, it is unreasonable to combine all the

results into one metric, as you have done.

 

 

 

Mike Lerer

Box 636

Londonderry, NH 03053

Home: 603-434-4205

Cell: 603-548-3704

-----Original Message-----

From: owner-stds-802-3-blade@listserv.ieee.org

[mailto:owner-stds-802-3-blade@listserv.ieee.org] On Behalf Of Altmann,

Michael W

Sent: Friday, October 29, 2004 1:44 AM

To: STDS-802-3-BLADE@listserv.ieee.org

Subject: [BP] Signaling spread sheet

 

Dear IEEE802.3ap TF Members (Signalig ad hoc),

 

Further to our discussions regarding comparison metrics for signaling

schemes, I created the first cut of a spreadsheet for coding selection

similar to that used by the 10Gbase-T Task force.  Comments/critiques

are all welcome.  I am not planning to discuss this in the signaling ad

hoc on 29 October, given it's relative new-ness.

 

Regards.

 

.../Mike

 <<IEEE Coding Table v3.xls>>