| Thread Links | Date Links | ||||
|---|---|---|---|---|---|
| Thread Prev | Thread Next | Thread Index | Date Prev | Date Next | Date Index |
Dear all,  We are writing to answer in detail some of the questions received during the presentation of 20/0659r3 TDM Multilink Operation and its associated SP (https://mentor.ieee.org/802.11/dcn/20/11-20-0659-03-00be-tdm-multilink-operation.pptx). Based on the discussion, we can also try to modify the SP in the interest of making it more agreeable to some delegates.  Below is a list of questions we recollect and responses for them. Please add any others that we have missed or any new ones.  ·      What is the difference between Enhanced Multi-Link Single Radio Operation (562r5) and our proposal 0659r3 (Yonggang/Rojan): o  562r5 considers a scheme where the performance of a single radio device can be enhanced via listening on 2 links and data tx/rx dynamically on any 1 of the 2 links at a time. o  659r3 has the same basic principle as 562r5, but specifies a capability that can be scaled up in terms of the number of links that it can support. So, if a device is single radio, 659r3 enables a mode of operation similar to 562r5. But if a device is multi-radio, 659r3 scales to enable higher order ML operation. For example, a device that has 2 nSTR radios in 5G/6G can use this scheme to listen on 4 links and do data Tx/Rx on 2 links at a time and perform closer to 4-link nSTR ML. Similarly, a device that has 2 STR radios in 5G/6G can scale up to perform similar to a 4-link ML device where there are 2 sets of links that are STR between each set but nSTR within each set. Looked at this way, 659r3 is a capability enhancing procedure that can be used irrespective of the underlying radio architecture even without requiring any additional hardware/software module.  ·      On TDM ML not providing any benefit compared to nSTR ML (Ming/Ross): You had earlier opposed 200/562r3 on the grounds that it requires a lower capability baseband in addition to a single radio and hence does not provide any cost/complexity advantage relative to a conventional dual-radio ML. So, please note the following with regard to 659r3: It provides the performance benefit of a higher order ML (up to 2M-link ML) for a conventional M-link ML device without requiring any additional hardware/software module. So, there is a clear performance benefit without any additional cost. Can you please revisit your objection in light of this observation and update?  ·      Removing RTS from examples of initial control messages and keep only MU-RTS (George): The most important factor here is the allowance of a switching delay required to switch from listen mode to operating mode, that the non-AP can indicate to an AP while setting up a TDM ML configuration. The purpose of the indication is that the AP admits such a TDM ML configuration only if it guarantees to provide for the indicated switching delay in a standards compliant manner. Both RTS and MU-RTS can be examples of initial messages that provide for such a delay, especially depending upon the range of the delay. For example, a switching delay of 10-15us can be met using RTS whereas that of 20-30us may need MU-RTS. As long as the AP agrees to grant the delay allowance in a standards compliant manner, shouldnâ??t it be the APâ??s choice how it provides for the delay, whether via RTS or MU-RTS or by some other means? From this perspective, we have listed RTS and MU-RTS as two possible examples.  ·      Removing the condition M=1, N=1 (George and others): We can remove this condition since it is already included in another passed SP (1943r8). We can do this by specifying (M>=1, N>1)    Regards, Sindhu   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 |