Thread Links | Date Links | ||||
---|---|---|---|---|---|
Thread Prev | Thread Next | Thread Index | Date Prev | Date Next | Date Index |
Hi all, I uploaded a modified SP3 (below) to better gauge the group interest in the case of AP MLD requesting a client to delete one of its links to the AP MLD. Specifically, the
“Yes” option has been split into AP recommending only (client can decline), or AP recommending as well as notifying (which the client cannot decline). My own position is the first option (i.e., what we had before, original SP3) - BSS optimization and traffic engineering need network-level knowledge and AP involvement, and we can think of additional conditions
later in the text that could help establish trust in the AP operation (a simple one Tomo pointed to for example is the AP not to the delete the last link standing in a link set). Please review and form an opinion for a possible straw poll in TGbe MAC call tomorrow/Thursday. Thanks, -----
Note: AP MLD practically can stop serving a non-AP MLD on a given MLO link; the mandate option (notification pointing to future removal time) aims to make the transition smooth by making
the non-AP MLD aware of the impending link removal. This is similar in nature to the “Disassociation Imminent” indication in the BTM Request frame.
From:
Payam Torab <torab@xxxxxxxx> Hi Arik, some of your comments are not aligned with 802.1 operation and ecosystem. What you see is a baseline formulated by AP and non-AP community. I would hope that you can support this baseline and then separately propose additional pieces that you see fit.
Regards, Payam
From:
Arik Klein <arik.klein@xxxxxxxxxx> Hi Payam,
See my further comments inline
Regards, Arik
From: Payam Torab <torab@xxxxxxxx>
Hi Arik,
If a non-AP tells AP it doesn’t want a link (from the MLO link set that it set up during association), for any reason, it’s free money (resource) returned to AP with no fight. Non-AP shouldn’t ask for this if it has any intention of using that link again, but if it does, well, it doesn’t want it. Smaller IEs, fewer transmissions, coming with no arbitration.
Please note the concept of “link” applies to non-AP/client. Cases designated as AP-initiated are in the form of request to non-AP to add/remove its link. On the AP MLD side, it can add and remove AP STAs serving the whole network. With that said,
[Arik Klein] I think that from system-wise perspective, such a “claim” that the non-AP STA is entitled to reject the AP MLD request to remove a link just because this resource has already been allocated for it is inefficient. I think the AP MLD has much broader perspective for the needs of all non-AP MLDs (due to dynamic changes in the traffic / network conditions), thus its request to remove a link with a specific non-AP MLD shall no be rejected (at least as long as we requires the AP MLD not to reject similar request from the non-AP MLD)
[Arik Klein] Agree with your point, but why shouldn’t it be “reciprocal” i.e. the non-AP MLD shall have courtesy indication for the time that the link will be
removed (as an option)?
Regards, Payam
From:
Arik Klein <arik.klein@xxxxxxxxxx> Hi Payam,
Thanks for sharing the presentation.
I have questions regarding proposed SP1 and SP3
In the proposed Straw Poll 1, when the non-AP MLD requests the associated AP MLD to remove a link – the AP MLD can’t reject this request.
On the contrary, in proposed Straw Poll 3, when the AP MLD requests the associated non-AP MLD to remove a link you propose either of 2 options:
My questions are:
Thanks for the swift reply.
Regards, Arik
From: Payam Torab <torab@xxxxxxxx>
Hello all, we will likely have time in our next MAC call to run straw polls for the ML reconfiguration topic I presented today. It would be nice if you could share your comments and questions here.
Regards, Payam 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 |