IEEE 802.3ap: ad-hoc signaling task force: August 5, 2004 meeting

Attendance:

Michael Altmann [Intel]
Stephen Anderson [Xilinx]
Majid Barazande-Pour [Vitesse]
Michael Beck [Alcatel]
Brian Brunn [Xilinx]
John D'Ambrosia [Tyco]
Doug Day [Toshiba]
Graeme Boyd [PMC-Sierra]
Adam Healey [Agere Systems]
Bill Hoppin [Synopsys]
Glen Koziuk [Vitesse]
Mary Mandich [Lucent]
David McCallum [Molex]
Roland Moubarak [TI]
Pravin Patel [IBM]
Petre Popescu [Quake]
Jimmy Sheffield [Tyco]
Jeffrey Sinsky [Lucent]
Fulvio Spagna [Intel]
Schelto	van Doorn [Intel]


    a.. Role call
    b.. Adam Healey [Agere] introduced the scope of the 'ad hoc' task force: a forum to discuss signaling issues. This is not where a decision will be made but, rather, the place where a framework is established to compare different signaling proposals.
    c.. Adam appointed Mike Altmann  [Intel] as 'ad-hoc' chair.
    d.. The following are comments to the presentation given by Mike Altmann [Intel]:
      a.. Slide1
        a.. Petre[Quake]: the most important thing is the  CRITERIA by which we evaluate the results (BER, latency, etc). We should strive for a common platform for objectives rather than a common platform for sims.
        b.. John [Tyco]: expressed concern regarding the two other PHYs and whether they should also be included in the list of objectives for this group
        c.. Mike[Intel]: .. if the group feels it's needed we will do that. It is certainly within the scope of the group but we must agree to do so.
        d.. Jeff[Lucent]: methodology is more important that simulation platform
        e.. Mike[Intel]: focus on methodology rather than simulation details
      b.. Slide 3
        a.. Bill[ ?]: presentations discussed in this forum should articulate test patterns, equalization details etc.
        b.. Mike[Intel]: we should definitely agree on test patterns. As an example, the patterns used for crosstalk evaluation
      c.. Slide 4 (methodology for comparing solutions)
        a.. Mary[Lucent]: would like to add power as a factor
        b.. ? : .. and cost ..
        c.. John[Tyco]: results should span a variety of channels not just the template
        d.. Mike[Intel] we need to define the strategy: worst case only, best/worst, best/intermediate/worst etc....
        e.. John[Tyco]: ..yes, and extend this to include manufacturing variability;
        f.. Mike[Intel]: how do we do this?
        g.. John[Tyco]: a sync-up ( with the channel ad-hoc) is required ... but we could include data from different builds
        h.. Petre[Quake]: agreed
        i.. Jeff[Lucent]: when comparing solutions we should keep in mind the finite precision of the equalizers  ...
        j.. Mike[Intel]: in the comparison of solutions for a signaling/equalization proposal (the two must go together), we will need to understand minimum requirements for tap coeff size, filter length etc etc.
        k.. Chair will try to capture discussion and send it out.
        l.. There is a need do to have more discussion on methodology guidelines (which should be reviewed at next conference).
        m.. Adam [Agere]: take advantage of the reflector for advance discussion.
        n.. Schelto[Intel]: Let's add error robustness (i.e. how easy it is to detect errors) to the list of parameters needed to compare different solutions 
    e.. Adam[Agere] reiterated that should craft link modeling discussion/guidelines so as not to exclude GB PHY or quad PHY
    f.. Meeting closed.