Hello all,
I deferred running the straw poll in the TGbe call yesterday, and I am initiating this email thread to allow for further discussions.
Several discussions with members have already happened offline, but we have been unable to fully agree on the merits of my proposal to allow a non-AP MLD to enable EMLSR but keep STAs in their existing PM state (awake or doze).
I see the current specified behaviour as a good optimization and I support efforts to enhance MLO with signalling which allows PM state changes across multiple links. However, given that EMLSR can systematically end up using only one of
the links for all UL and DL between an AP MLD and an non-AP MLD, I believe the additional power consumption and the complexity in the implementation at the non-AP MLD (to try and get the PM=1 frame out on a link which struggles to get a TXOP) calls for a backup
solution to allow for legacy link behaviour (no implicit change of PM state).
The proposed change allows keeping the existing behaviour for the non-AP MLD, and requires from the AP MLD to support a EMLSR enablement which keeps the non-AP links in their existing PM states. I claim that this would be a minor ask from
the AP MLD, as it is standard behaviour at all times except during EMLSR enablement.
All comments, suggestions and questions are very welcome.
Kind regards,
Michail
|
|
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