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

Re: [STDS-802-11-TGBN] PDT-MAC-NPCA

Hi Matt,

 

Thanks very much for the clarification! Two more here:

 

For the “TXTIME” stuff, I agree with you that it’s complicated to make the change and the calculation. My suggestion is making it simple for now and leave that for later discussion, possibly involve PHY guys.  So, the text could be something like: 

the duration (TBD for how to determine the value) of the received PPDU is greater than the NPCA value indicated in the most recently received or transmitted NPCA Minimum Duration Threshold field corresponding to its BSS.

 

Another issue is about untriggered UL transmissions mode.

The SFD says “Whether the mode is for all associated non-APs or per non-AP is TBD”. But r9 writes “If the STA is a non-AP STA and the associated AP has disabled the use of untriggered UL transmissions on the NPCA primary channel, then the non-AP STA shall not initiate a TXOP on the NPCA primary channel.” Sounds like r9 making it a mode for all associated non-APs. I’d suggest the following change:

 

A non-AP STA that supports NPCA operation shall announce its NPCA Switching Delay and NPCA Switch Back Delay in TBD frames.

Add the following:

An NPCA AP may enable or disable the use of untriggered UL transmissions on the NPCA primary channel, by transmitting TBD frames.

i.               Whether the mode is for all associated non-APs or per non-AP is TBD.

ii.              TBD whether MU EDCA parameters mechanism and or some other mechanism is used to disable untriggered UL transmissions on the NPCA primary channel.

 

c.   If the STA is a non-AP STA and the associated AP has disabled the use of untriggered UL transmissions on the NPCA primary channel, then the non-AP STA shall not initiate a TXOP on the NPCA primary channel.

Remove the following:

i.    TBD whether MU EDCA parameters mechanism and or some other mechanism is used to disable untriggered UL transmissions on the NPCA primary channel.

 

BR,

Chaoming

发件人: Matthew Fischer <matthew.fischer@xxxxxxxxx>
发送时间: 2024年12月10日 4:08
收件人: STDS-802-11-TGBN@xxxxxxxxxxxxxxxxx
主题: Re: [STDS-802-11-TGBN] PDT-MAC-NPCA

 

 

1762r9 now available:

 

asdf

 

Thank you for the comments, see some responses inline:

 

 

On Sun, Dec 8, 2024 at 10:52 PM 罗朝明(Chaoming Luo) <luochaoming@xxxxxxxx> wrote:

Hi Matt,

 

Thanks for the update. Still there are points in r7 not comply with SFD:

 

In SFD “due to OBSS traffic or other TBD conditions”,  meaning “other TDB conditions” are NOT necessarily be sub-category of “OBSS traffic. However, r7 writes TDB conditions only in sub-items of OBSS PPDU or exchange.

 

Updated to include a new separate item c) for TBD conditions

 

 

In SFD, no agreement on which exact subchannel will be selected for NPCA p-channel. However, r7 limits NPCA p-channel to be in S160 or S80 or S40. Which may not be a reasonable limitation. E.g., when BSS A of 320 MHz overlaps with BSS B of 80 MHz, why A can not choose the S80 of the P160 as its NPCA p-channel?

Paragraph deleted

 

 

PLME-TXTIME.confirm primitive is generated in the transmitter side but not receiver side, so the duration of the received PPDU can’t be determined by the receiver by using TXTIME.  PSDU_LENGTH in RXVECTOR may help on this.

 

PSDU_LENGTH is octets, not time, so it will not give the answer that the MAC needs.

The MAC needs many more parameters than just PSDU_LENGTH to calculate a PPDU duration.

And once it has all of the parameters it needs, the equation to determine the PPDU duration is very complex.

In the case of TX PPDU, there are quite a few separate TXTIME equations that are in the baseline, and by using PLME-TXTIME.request, the complexity of choosing the correct parameters and the correct TXTIME equation are all invisible to the MAC.

 

So exactly the same complicated determination of a PPDU duration would exist for the MAC if it wants to calculate a duration for an RX PPDU, hence, the simplest solution would seem to be for the MAC to just use the same PLME-TXTIME.request, but instead of passing TXVECTOR, pass an RXVECTOR.

 

But, there are issues with this:

 

If the MAC uses an RXVECTOR in the TXTIME.request, the PHY should then return the TXTIME for the RX PPDU.

Theoretically.

Of course, it is much more complicated than this.

 

Even though TXVECTOR and RXVECTOR are defined with the SAME TABLEs, they are actually different because many of the parameters in the table are unique to RX vs TX and vice versa.

So it is possible that some of the parameters needed are missing in some RXVECTORS (e.g. the L_LENGTH is not included in the RXVECTOR for an HE PPDU!)

And a PHY would then know that there is something wrong if it gets a PLME-TXTIME.request(RXVECTOR)

 

So the simplest fix for this might be to add a PPDU_DURATION parameter to RXVECTOR for ALL FORMAT values.

Then we would still have to add language to the PHY to tell it how to use TXTIME equations with the RX PPDU parameters to create this value.

I.e. the PHY would pass this value in the RXVECTOR.

 

There are still possibly some additional complications, as the PHY could compute a PPDU duration when it has received the L_SIG but before it has received a HE header, for example and then another one after it receives another PHY header. So the PHY might need to pass more than one RXVECTOR (it already does, because it passes one RXVECTOR at PHY header successful decode (RXSTART) and another one at the end of the PPDU reception in the RXEND)

 

I am not certain that I want to start making such changes until we have an agreement as to exactly what the best solution is.

And I hope that the current language is good enough for now.

 

I believe that this sort of detail is best left to a later date.

 

 

 

BR,

Chaoming

发件人: Matthew Fischer <matthew.fischer@xxxxxxxxx>
发送时间: 2024年12月7日 6:48
收件人: STDS-802-11-TGBN@xxxxxxxxxxxxxxxxx
主题: Re: [STDS-802-11-TGBN] PDT-MAC-NPCA

 

 

1762r7 PDT MAC NPCA is on the server:

 

 

This version includes many editorial changes all of which were prompted by comments received from Mark Rison.

 

 

 

On Fri, Dec 6, 2024 at 2:50 AM Mark Rison <m.rison@xxxxxxxxxxx> wrote:

Thanks for this, Matt.  I attach some comments.

 

Mark

 

--

Mark RISON, Standards Architect, WLAN   English/Esperanto/Français

Samsung Cambridge Solution Centre       Tel: +44 1223  434600

1 Cambridge Square, Cambridge CB4 0AE   Fax: +44 1223  434601

ROYAUME UNI                             WWW: http://www.samsung.com/uk

 

From: Matthew Fischer <matthew.fischer@xxxxxxxxx>
Sent: Thursday, 5 December 2024 20:27
To: STDS-802-11-TGBN@xxxxxxxxxxxxxxxxx
Subject: Re: [STDS-802-11-TGBN] [EXTERNAL] [STDS-802-11-TGBN] PDT-MAC-NPCA

 

 

Thank you for all of the comments received during the call and through the reflector.

 

I have attempted to address each of the comments in the latest revision that is now available on the server:

 

 

In addition to the revision notes, you will find the following in the new DISCUSSION section:

 

The use of the term “received PPDU” in the descriptive behavioural subclauses (e.g. 37.x NPCA) is problematic because the OBSS PPDU that is used to trigger a switch to an NPCA operation on a subchannel of the BSS is not always actually “received”. For the Control frame cases, the PPDU(s) are received, but for the PPDU without a preceding Control frame, only the PHY header of the OBSS PPDU is received before the switch occurs. So for that case, we cannot use “received”. The MAC does receive information from the PHY after a valid PHY header reception, in the PHY-RXSTART.indication() primitive and this does contain an RXVECTOR.

 

Text has been modified appropriately.

 

 

On Thu, Dec 5, 2024 at 7:44 AM Brian Hart (brianh) <brianh@xxxxxxxxx> wrote:

Hi Matt

 

As per thread in chat, with a little augmentation:

 

… (is the field really called just "TXOP"?)

… I think they changed it to be TXOP … Take it back, it is still called TXOP_DURATION

… this seems like the TXVECTOR or RXVECTOR parameter. The subfield is still called "TXOP"

… Since this is MAC text, the text should refer to the parameter used in the MAC-PHY interface (“TXOP_DURATION parameter in the TXVECTOR”), not something below that interface (“TXOP field”)

 

Cheers

Brian

 

From: Matthew Fischer <matthew.fischer@xxxxxxxxx>
Sent: Tuesday, December 3, 2024 10:21 AM
To: STDS-802-11-TGBN@xxxxxxxxxxxxxxxxx
Subject: Re: [STDS-802-11-TGBN] [EXTERNAL] [STDS-802-11-TGBN] PDT-MAC-NPCA

 

 

There will be no r5 at this time.

 

On Tue, Dec 3, 2024 at 10:19 AM Matthew Fischer <matthew.fischer@xxxxxxxxx> wrote:

 

John,

 

If you retain the term NPCA, and then use the term Alternate Primary Channel, then you lose a nominal coupling to the NPCA mechanism.

 

If we change the name of NPCA to TACA = Temporary Alternate Channel Access, and then call the new primary as the Alternate Primary Channel, then there is a nice nominal pairing between the name of the feature/mechanism/operational mode and one of the items that is a subject of that mode.

 

e.g. If we use TACA, then when a STA switches to TACA operation, the STA is performing Temporary Alternate Channel Access on an Alternate Channel where there is an Alternate Primary Channel.

 

You could even drop the T and just have ACA

Alternate Channel Access

Alternate Channel

Alternate Primary Channel

 

Some might argue for changing "Channel" to "Subchannel", as the Alternate Channel is supposed to be a portion (i.e. subchannel) of the normal BSS Operating Channel

 

Bur dropping Temporary loses the ephemeral nature of the channel, hence my inclusion of "temporary", so, I think it is better to drop the Alternate label, as Temporary sort of covers the alternate nature of it anyway:

 

Temporary Subchannel Access = TSA

Temporary Subchannel

Temporary Primary Channel

 

Or, you could use Interim:

 

Interim Subchannel Access

Interim Subchannel

Interim Primary Channel

 

more choices:

 

Transient

Ephemeral

Fugacious

Capricious

 

 

 

 

On Mon, Dec 2, 2024 at 10:11 AM Wullert, John R II (PERATON LABS) <jwullert@xxxxxxxxxxxxxxx> wrote:

Matt,

  One note – when I proposed the phrase “Alternate Primary Channel Access”, I was not proposing that we use “APCA primary channel”, but that we use “Alternate Primary Channel” to refer to the channel used for access control.

John

 

From: Matthew Fischer <matthew.fischer@xxxxxxxxx>
Sent: Wednesday, November 27, 2024 3:13 PM
To: STDS-802-11-TGBN@xxxxxxxxxxxxxxxxx
Subject: Re: [STDS-802-11-TGBN] [EXTERNAL] [STDS-802-11-TGBN] PDT-MAC-NPCA

 

 

After receiving some editorial comments regarding 1762r2, I have made many of the suggested editorial changes and then added some of my own.

The result is a new r3 version that is now available on the server:

 

 

 

 

 

On Tue, Nov 26, 2024 at 3:09 PM Matthew Fischer <00000959766b2ff5-dmarc-request@xxxxxxxxxxxxxxxxx> wrote:


Thanks to those who submitted comments on 1762r1.

Due to the large volume of motions that were passed regarding the NPCA feature, there is a new revision of 1762 that has been uploaded.

This revision is constructed from several comments/proposed text sections that I received from various TTT members and then consolidated into the r2 version of 1762.

The changes are vast.

This latest revision can be found on the server:

 

 

Note that because of the amount of changes made to the earlier revision of the document, those comments that were sent to this email thread were not directly addressed.

 

As to the comment from John Wullert regarding terminology, I agree that the term NPCA primary channel seems a bit clumsy upon expansion of the acronym.

John's suggestion is an interesting alternative, but I would not unilaterally adopt it without first soliciting additional opinion on the matter.

[of course, at this point, no one else has responded... so maybe there is no other opinion!]

[and I'll note that if you do not expand the acronym, the term isn't so clumsy]

[and I'll note that even with John's proposal, upon expansion, you still get a "double primary" problem - e.g. APCA Primary Channel = Alternate Primary Channel Access Primary Channel]

 

How about TACA - Temporary Alternate Channel Access

 

No change has yet been made to that term.

 

 

 

Matthew Fischer

Nice Guy

Broadcom Inc.

+1 408 543 3370 office

 

 

On Mon, Nov 25, 2024 at 8:45 AM Wullert, John R II (PERATON LABS) <jwullert@xxxxxxxxxxxxxxx> wrote:

Matt,

   Thanks for this.  The one comment I have is related to terminology – It seems to me that “non-primary channel access primary channel”, the expansion of “NPCA primary channel”, has the potential to lead to confusion.  It might be better to call this feature “Alternate Primary Channel Access”?  In that case, what is currently labeled “NPCA primary channel” would become the “alternate primary channel”, which seems much more descriptive and clearer to me.  (Given that the motions use NPCA, it may be that this is the wrong place to have the discussion…)

John

 

 

 

 


 

--

Matthew Fischer

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


 

--

Matthew Fischer


 

--

Matthew Fischer

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


 

--

Matthew Fischer

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

图像已被发件人删除。

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


 

--

Matthew Fischer

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


 

--

Matthew Fischer

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

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