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

[STDS-802-11-ARC] Update to proposed liaison to TGbn on L4S



--- This message came from the IEEE 802.11 ARC Reflector ---

All,

 

On today’s call, we reviewed the proposed liaison document to TGbn, with suggested responses to the CIDs they received about using L4S on non-AP STAs.  The updated draft can be found here: https://mentor.ieee.org/802.11/dcn/26/11-26-0755-01-0arc-liaison-to-tgbn-on-l4s-comments.docx

 

The changes were:

  • Removed CIDs and definition/acronym changes not related to non-AP STA operation
  • Removed subclause 5.2.4, which is unchanged
  • Modified Annex <x>.2 text per suggestion
  • Removed Annex <x>.3. 
  • Also rewrote the “fairness” bullet in the Discussion section.

 

To that last point, I would request review by everyone of the new text in the Discussion section (I’ve copied it here, below).  I tried to capture the tone and feel of the text direction as discussed on the call.  I am very open to suggestions for any wording changes.

 

To re-iterate the plan: The hope is to finalize the draft here on the reflector, and come into the Antwerp F2F session ready to approve a motion to send this to TGbn, in our first meeting on Monday morning.  So, please express any comments/concerns here, so we can try to converge before the F2F.

 

Thanks.  Mark

---

Text from Discussion (updated):

 

  1. For a non-AP STA that does happen to be contained in the device that is the source of a data flow that results in congestion, an implementation consideration should be made for “fairness” in relieving the congestion.  While a “local back pressure” method can result in very quick reaction to congestion – which would seem desirable – it can also result in the first device to detect congestion taking all the burden of relieving the congestion, rather than allowing the larger system to react consistently across all participating devices.  To get better fairness, the device could delay the local back-pressure reaction, so that other STAs in the BSS also notice the congestion and take similar action to slow flows, but this relies on consistency in the implementation across all the devices, and guesswork on how long the delay should be.  One way to enable standardized fairness among the non-AP STAs in a BSS is for the STAs to detect, signal, and react to congestion similarly to each other, by using L4S signaling on all the STAs.  Thus, all the devices in the BSS can equivalently detect and report the congestion, and the mechanisms for resolution can be evenly spread across all the streams and devices involved in the congestion.  This option can be made available to implementation choice, by including support for the (optional) use of L4S signaling at non-AP STAs, even when they are the source of the flow(s).

 


To unsubscribe from the STDS-802-11-ARC list, click the following link: https://listserv.ieee.org/cgi-bin/wa?SUBED1=STDS-802-11-ARC&A=1