[RPRWG] Minutes of 2003Jun25 Fairness AdHoc meeting

["David V James" <dvj@xxxxxxxxxxxx>]

All,

Minutes follow.
Please not that DVJ will be gone and out-of-contact
till 2003Jul15.

DVJ

David V. James
3180 South Ct
Palo Alto, CA 94306
Home: +1.650.494.0926
      +1.650.856.9801
Cell: +1.650.954.6906
Fax:  +1.360.242.5508
Base: dvj@xxxxxxxxxxxx

**** Minutes of Fairness AdHoc ****

Todays meeting:
Date: Wednesday, June 25, 2003 
Start Time: 10:00:00 AM Pacific Daylight Time 
End Time: 12:55:00 PM Pacific Daylight Time 
Parties: 15 
Dial-in Number: 1-702-835-5000 (Las Vegas, Nevada) 
Participant Access Code: 8021725
Attending: 
   Harry Peng:         hpeng@xxxxxxxxxxxxxxxxxx
   David James:        dvj@xxxxxxxxxxxx
   Jon Schuringa:      jon.schuringa@xxxxxxxxxxxx
   Stein Gjessing:     steing@xxxxxxxxx
   John Lemon:         JLemon@xxxxxxxxxxxx

Subscriber list:
   Khaled Amer:        amer@xxxxxxxxxxx
   Leon Bruckman       leonb@xxxxxxxxxxxxx
   Stein Gjessing:     steing@xxxxxxxxx
   David James:        dvj@xxxxxxxxxxxx
   Kshitij Kumar       kkumar@xxxxxxxxxxxxxx
   John Lemon:         JLemon@xxxxxxxxxxxx
   Robert D. Love:     rdlove@xxxxxxxx
   Paritosh Kulkarni:  paritosh@xxxxxxxxxxxxxxx
   Harry Peng:         hpeng@xxxxxxxxxxxxxxxxxx
   K. K. Ramakrishnan  kkrama@xxxxxxxxxxxxxxxx
   Yan Robichaud       yan@xxxxxxxxxxxxxxx
   Jon Schuringa:      jon.schuringa@xxxxxxxxxxxx
   Necdet Uzun:        nuzun@xxxxxxxxx
   Dongmei Wang        mei@xxxxxxxxxxxxxxxx


******** NEXT MEETING DATE ********
Date: Wednesday, July 02, 2003 
Start Time: 10:00:00 AM Pacific Daylight Time 
End Time: 12:55:00 PM Pacific Daylight Time 
Parties: 15 
Dial-in Number: 1-702-835-5000 (Las Vegas, Nevada) 
Participant Access Code: 8021702 
 

Minutes of meeting, 2003Jun25:
  1) Meeting times
     Short presentations Monday AM
     Evening ad-hoc on site
     DVJ: Ask Mike for times for these
     Contributions due by Wednesday noon 2003Jun16
     Ship Monday presentation by Friday noon 2003Jun18,
      comments back by Sunday 20:00 PST

  2) Document review
     Question that may require more thought:
       Is there any problem if downstream shaper applies to
        only classB/C in STQ, not the PTQ?
       The subclassA1 in the PTQ may be a concern.
       Stein: A1 should behave as currently
       DVJ: A1 would behave equivalently to A0
     Other solutions: no silicon change (Harry Peng)

     Discussion if S61 were to be large, does that solve problem?
     * If LRTT remain dominant, then low-threshold has to be a
       smaller percentage, because the upstream contributors
       cumulative problem also becomes worse.
     * If LRTT became much smaller than the STQ low-threshold,
       then fairness would have a chance to kick in.
     * (DVJ) But how long does fairness take to kick in?
       This would become a fairness response-time question, dependent
        on the time to stop fairness-elibible upstream sources.
       Note that stopping fairness-elibible sources is required and
        helps, but could be insufficient. This still leaves
        open the question of upstream classB sources.

  3) Jon concerns presented:
     The downstream shaper keeps the STQ from ever filling.
     This would effectively disable all fairness, since the
       low-threshold would never be reached.
     This would also be solved by same solution of having
       the STQ also be limited by the downstream shaper.

  4) Actions to be done on simulations:
     Preliminary details:
       2,000km S30-to-S31 distance
       STQ buffer-time corresponds 1% of that distance
       1Gbs link rate
       Aggressive fairness model
     1Km for each other stations distance
       Harry Peng
       Stein Gjessing
       Jon Schuringa
     Action:
       Harry to send email on scenario details, including
       STQ size details in bytes

  5) ClassB expectations.
     John Lemon, Harry, DVJ, Stein:
       The only classB limit should be approximately:
       * dual-queue: classB= (line-classA)
       * single-queue:
         classB= (lineRate-classA)*fairnessThreshold
           where fairnessThreshold is somewhere between
             rateLowThreshold and rateHighThreshold (90-95%)
     Jitter bound is on the order of FRTT time
       We don't want worst case to become larger
       We don't require that worst-case be as small as classA

  6) No silicon changes
     Harry emphasized hardware is underway
     Some other expressed concerns that breaks must be fixed


Action items:
  1) Stein and DVJ (revised downstream shaper proposal&eval):
     If downstream does effect STQ, think about how classA1
     is affected (Stein, DVJ, and option options).
     Also, does downstream shaper need to effect classA1 add.
  2) Harry Peng:
     Simulation details to be sent to other
  3) Simulations by:
     Harry Peng
     Stein Gjessing
     Jon Schuringa
  4) Normalized fairRate: mention in day, talk in evening,
     ad-hoc to have preview along with presentation material.
  5) Meeting for next week, DVJ to do FreeConference and
     post in the minutes.
  6) Host is TBD.
    

DVJ