| Thread Links | Date Links | ||||
|---|---|---|---|---|---|
| Thread Prev | Thread Next | Thread Index | Date Prev | Date Next | Date Index |
Hi Stephen,
The current EMLSR specification does not preclude a non-AP MLD in EMLSR mode from listening on one link during frame exchanges on any other link. This is a matter of capability of the non-AP MLD architecture that implements the EMLSR mode.
This is noted in multiple sections of the specification:
“When a non-AP MLD is operating in the EMLSR mode, a STA affiliated with a non-AP MLD that is operating on one of the EMLSR links is considered to have lost medium synchronization if it is not able to perform CCA during frame exchanges that includes the link switch delays between an AP affiliated with an AP MLD and one of the other STAs operating on the other EMLSR links, which are affiliated with the same non-AP MLD.”
>> The phrase “if it is not able to perform CCA during frame exchanges” means that it is not always certain that a non-AP MLD in EMLSR mode is not able to perform CCA on other links during frame exchanges on one link. The possibility of being able to perform CCA is left open.
“The non-AP MLD shall be switched back to the listening operation on the EMLSR links after the time indicated in the EMLSR Transition Delay subfield”
>>This specifies the return to listening operation for the entire non-AP MLD i.e. return to listening operation on all the EMLSR links including the one on which the frame exchange was performed. It too does not leave out the possibility of the non-AP being able to listen on the other EMLSR links during frame exchanges on any one link. This is a matter of capability. Some architectures may not be able to listen (and the specification allows that) while some other architectures may be able to listen (and the specification does not preclude that).
Regards,
Shubho
On Fri, 9 Sept 2022 at 20:36, Stephen McCann <mccann.stephen@xxxxxxxxx> wrote:Shubho,thanks for the feedback. I think the term "switching" is justified, as the non-AP MLD listens on multiple receive chains and then "switches" its operation to exchange frames on one link. I think the term "perform" implies that the non-AP MLD may still be listening on the other receive chains.However, I must admit that finding the correct wording for this definition has been tough !With regard to section 35.3.17, I appreciate what you mean, but I'm trying to avoid either referencing or quoting normative text within the definition itself, as I don't believe that this provides a clear definition of what EMLSR operation actually is.Kind regardsStephenOn Fri, 9 Sept 2022 at 05:52, Shubhodeep Adhikari <00000c144a46bcee-dmarc-request@xxxxxxxxxxxxxxxxx> wrote:Hi Stephen,
Thanks for the contribution 1196r3.
I have a comment on the proposed new definition of EMLSR “A mode of operation that allows a non-access point (non-AP) multi-link device (MLD) with multiple receive chains to listen on a set of enabled links, then switch to one available link within the set, for all of the following frame exchanges. mode of operation that allows a non-access point (non-AP) multi-link device (MLD) with multiple receive chains to listen on a set of enabled links when the corresponding stations (STAs) affiliated with the non-AP MLD are in awake state for an initial Control frame sent by an AP affiliated with an AP MLD in a non-high-throughput (non-HT) (duplicate) physical layer (PHY) protocol data unit (PPDU) with one spatial stream, followed by frame exchanges on the link on which the initial Control frame was received”
My concern is that it tries to define EMLSR operation of listening on the set of links and then “switching” to one of them which entails then the question of switching back, etc. How about an alternate definition as below?
Enhanced multi-link single radio (EMLSR) operation: A mode of operation that allows a non-access point (non-AP) multi-link device (MLD) with multiple receive chains to listen on a set of enabled links called EMLSR links but perform frame exchanges on only one of those links at any time.
A more complete definition requires mentioning other aspects, such as what is already mentioned at the start of section 35.3.17.
Regards,
Shubho
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