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

Re: [STDS-802-11-TGBE] Group addressed frames delivery discussion

Hi Po-Kai, Docun,

 

I agree the intention to address the missing and duplicated issue during the indicated link switch process. But I’m concern how to implement it according to current SPEC?

As you know, it’s impossible to know whether there is any un-transmitted groupcast BUs on both links of AP MLD unless we define new signaling to indicate it. What we talked the concept of BUs and TIM/DTIM is still under the PS scheme, which is also aligned with what Docun mentioned that the service duration is behind DTIM.

I don’t have strong opinion to tie the link switch process to the non-AP MLD PS status. Seems we spend so many efforts to address such corner case.

I think it’s better to define a common behavior that AP MLD can indicate whether there is any groupcast BUs on any links via new signaling, no matter to what’s the non-AP MLD status, and non-AP MLD can switch the indicated link according to the signaling.

 

 

Thanks

 

Best Regards

 

Jay Yang

 

From: Huang, Po-kai <po-kai.huang@xxxxxxxxx>
Sent: 2021年1月14日 0:12
To: STDS-802-11-TGBE@xxxxxxxxxxxxxxxxx
Subject: [Suspected Marketing Mail] Re: [STDS-802-11-TGBE] Group addressed frames delivery discussion

 

Hi all,

 

              Just want to concur Duncan that we already have the following agreed motion to align the group addressed reception with baseline, and in baseline a STA already can not get out of a existing group addressed reception period, so all these directions align with what we have agreed.

 

 

  • Do you support the following in R1?
    • If a non-AP MLD intends to receive group addressed data frame, the non-AP MLD shall follow the baseline rules to receive the group address data frames on any one link that the non-AP MLD selects to receive group addressed data frames.
    • A group addressed data frame that is expected to be received by the non-AP MLD shall be scheduled for transmission in all the links setup with the non-AP MLD. 

Motion 144, #SP327

 

Best,

Po-Kai

 

From: Duncan Ho <dho@xxxxxxxxxxxxxxxx>
Sent: Wednesday, January 13, 2021 7:03 AM
To: STDS-802-11-TGBE@xxxxxxxxxxxxxxxxx
Subject: [STDS-802-11-TGBE] Group addressed frames delivery discussion

 

Hi all,

 

Thanks for the comments and discussion today regarding my contribution 20/442r4 Group Addressed Frame Delivery for EHT.

 

I ran out of time to explain more so let me try to do it here based on the passed discussion and assumptions. Please let me know if you have any questions or comments.

 

  • Desirable behavior/requirement: Non-AP MLD will be able to receive these frames on any links for full flexibility for the non-AP MLD
  • Desirable behavior/requirement: The non-AP MLD selects to receive these frames but does not need to tell the AP MLD
  • => Agreed solution/path: the group addressed frames are duplicated and transmitted on all the links (since the AP doesn’t know a prior which link the non-AP MLD selects)

 

  • However, since these frames use different SN space on different links, the non-AP MLD will not be able to do any duplicate detection on two frames received from two different links
    • This is NOT  a problem *IF* the non-AP MLD doesn’t change the selected link
    • The duplicates or miss occurs only if the non-AP MLD changes link so the question is how to solve these two problems. Frames can be missed or duplicated during the link switch
    • To solve the “miss” problem, the non-AP MLD should avoid switching link in the middle of a group addressed frames delivery  on the selected link (a train of frames on that link being delivered)
    • To solve the “duplicate” problem, we added cross-link BU indication bcast on each link. The non-AP MLD should avoid switching link if the AP on the selected link indicates there are still BUs on the other links (because it may receive duplicated frames undetected otherwise)
      • Note the above duplicate avoidance is not 100% but should work well in most cases (where the group addressed delivery frames come sporadically as bursts and get duplicated to all links around the same time)
      • To solve the duplicate problem 100%, one obvious way is to use a common SN for all the frames but that option was met with heavy objections

 

Thanks,

Duncan

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

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

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