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

[STDS-802-11-TGAX] 答复: [STDS-802-11-TGAX] 答���: [STDS-802-11-TGAX] Allowed puncturing patterns



Hi Mark,

 

Thanks for the reply.

 

Simply saying, Opt C is better than Opt B.

 

First of all, Opt B doesn’t support full P80 and may lose 20% Tput efficiency if full P80 is idle, which is a lotJ Nobody can deny that.

 

Moreover, when The Tx transmits full P80 by Opt C, what’s the loss when the Rx only receives HE-SIG-B by P40? It is not worse than Opt B. The Rx can never use S40 under Opt B anyway. I don’t understand why transmitting full P80 under Opt C has issues. The logic is not right. The reason why the Tx have duplicated contents in S40 is because of Opt C. For Opt B, the duplication never happens.

 

 

 

Regards

于健, Ross Yu

 

发件人: Mark RISON [mailto:m.rison@xxxxxxxxxxx]
发送时间: 2020513 20:38
收件人: Yujian (Ross Yu) <ross.yujian@xxxxxxxxxx>
抄送: STDS-802-11-TGAX@xxxxxxxxxxxxxxxxx
主题: Re: [STDS-802-11-TGAX] 答复: [STDS-802-11-TGAX] Allowed puncturing patterns

 

> You are correct, Opt C is my preference. And the current description in D6.0 makes people think it as either Opt B or Opt C. And I see no reasons to force the Tx to puncture one subchannel in P80, which is OptB.

> I encourage the group to move forward with Opt C. In that case, we can have a clear description about preamble puncture and with no technical changes at this stage.

 

The wording is ambiguous in D6.0 for "case 7", so it is far from clear that option C is "no technical changes";

for me option B looks like the "no technical changes" option:


"the primary 40 MHz channel in the primary 80 MHz channel is present" -- allows S40 to be present
"in the primary 80 MHz of the preamble only the primary 40 MHz is present" -- disallows S40 to be present
"in the primary 80 MHz of the preamble the primary 40 MHz is present, and at least one 20 MHz subchannel that is not in the primary 40 MHz is punctured" -- ambiguous; either requires one or both 20M in S40 to be absent, or allows S40 to be present if at least one 20M in S80 is absent
"the primary 40 MHz in the primary 80 MHz is present, and at least one 20 MHz subchannel that is not in the primary 40 MHz is punctured" -- ambiguous; either requires one or both 20M in S40 to be absent, or allows S40 to be present if at least one 20M in S80 is absent

 

> From implementation point of view, I should say even Opt B disallows full P80 transmission, the AP can still transmit full P80, and it is transparent to STAs.

 

The problem is that this impairs the receiver's sensitivity, because if it's

been told that the S40 might not be present, it will/might only look at the P40

and so not be able to make use of the duplicated content channels in the S40,

which are in fact actually there.  So in this respect the AP should not say

"case 7" if in fact the full P80 is present.

 

In the end, maybe it's a question of what the puncturing signalling is trying to

indicate:

 

a) Where the receiver needs to look to be sure to find content channels

 

or

 

b) That some 20M subchannels might or might not be missing

 

I think a) is more useful and important to a receiver than b), so I lean towards

option B, if I have to choose between that and option C.  But since I think there

are regulatory issues with multiple adjacent punctured channels, I lean towards

option A1, if I have to choose between that and option B.

 

Thanks,

 

Mark

 

--

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

Samsung Cambridge Solution Centre       Tel: +44 1223  434600

Innovation Park, Cambridge CB4 0DS      Fax: +44 1223  434601

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

 

On Wed, 13 May 2020 at 09:47, Yujian (Ross Yu) <ross.yujian@xxxxxxxxxx> wrote:

Hi Mark,

 

Very good job by summarizing all the possible way forward. Really appreciate that.

 

You are correct, Opt C is my preference. And the current description in D6.0 makes people think it as either Opt B or Opt C. And I see no reasons to force the Tx to puncture one subchannel in P80, which is OptB. From implementation point of view, I should say even Opt B disallows full P80 transmission, the AP can still transmit full P80, and it is transparent to STAs.

 

For Opt D, it changes the meaning of a non-punctured mode. I think it is too much change at this stage.

 

I encourage the group to move forward with Opt C. In that case, we can have a clear description about preamble puncture and with no technical changes at this stage.

 

All further comments are welcome.

 

regards

于健 Ross Yu

Huawei Technologies

 

发件人: *** 802.11 TGax - HEW - High Efficiency WLAN *** [mailto:STDS-802-11-TGAX@xxxxxxxxxxxxxxxxx] 代表 Mark RISON
发送时间: 2020512 15:27
收件人: STDS-802-11-TGAX@xxxxxxxxxxxxxxxxx
主题: Re: [STDS-802-11-TGAX] Allowed puncturing patterns

 

I'm mildly confused, because on re-reading SP4 I think it does not allow

the "xx-x" pattern in S80 someone mentioned last week (I forget whom

now); I really have to stretch the interpretation of the bullet to make

it possible.

 

Anyway, I've attempted to capture a smörgåsbord of puncturing options

for the group to consider.  I am only presenting the HE-SIG-A wording;

the ~46 other places in which the preamble puncturing options are described

would of course be aligned with this once we agree on something.

I am not showing change tracking w.r.t. D6.0 because I don’t think it’s

useful at this point anymore (it would be mostly underlines and strikethroughs!).

 

Option A: based on the SPs that had a reasonably conclusive outcome as captured in the agenda (20/0538r13), with my interpretation shown:

 

SP1) For 160M/80+80M, should having all the 20M subchannels in the secondary 80M channel punctured be allowed?

Y/N/A = 0/23/6 [my interpretation: no, not allowed]

SP3) For 160M/80+80M, should D6.0 be interpreted as allowing having no 20M subchannels in the primary 80 MHz channel punctured (only in the secondary 80 MHz channel)?

Y/N/A = 6/9/18 [my interpretation: no, nowhere near enough support to allow this]

SP4) Do you agree that for 160M/80+80M, allow only a maximum of two adjacent 20 MHz subchannels to be punctured?

       When puncturing 40 MHz in secondary 80, only puncturing either the lower 40 MHz or the upper 40 MHz

Y/N/A = 19/5/8 [my interpretation: yes, only allow max 2, but maybe two possible interpretations of bullet; see below]

 

Option A1: interpret bullet as saying “if you puncture two in the S80, it has to be either the lower 40 or the upper 40; you can’t puncture the other one at all”:

 

Table 27-20—HE-SIG-A field of an HE MU PPDU, Bandwidth field

 

Set to 4 for preamble puncturing in 80 MHz, where in the preamble the only punctured subchannel is the secondary 20 MHz channel.

Set to 5 for preamble puncturing in 80 MHz, where in the preamble the only punctured subchannel is one of the two 20 MHz subchannels in the secondary 40 MHz channel.

Set to 6 for preamble puncturing in 160 MHz or 80+80 MHz, where in the preamble the only punctured subchannels are the secondary 20 MHz channel and zero to two of the 20 MHz subchannels in the secondary 80 MHz channel.  If two of the 20 MHz subchannels in the secondary 80 MHz channel are punctured, these are either the lower two or the higher two.  No more than two adjacent 20 MHz subchannels are punctured across 160 MHz.

Set to 7 for preamble puncturing in 160 MHz or 80+80 MHz, where in the preamble the only punctured subchannels are one or both of the 20 MHz subchannels in the secondary 40 MHz channel and zero to two of the 20 MHz subchannels in the secondary 80 MHz channel.  If two of the 20 MHz subchannels in the secondary 80 MHz channel are punctured, these are either the lower two or the higher two.  No more than two adjacent 20 MHz subchannels are punctured across 160 MHz.

 

Option A2: interpret bullet as saying “if you puncture one of the 40s in the S80, you can’t entirely puncture the other one but you can puncture a 20 in it” (note: I don’t believe this is a valid interpretation, because we’d already agreed that we couldn’t puncture the entire S80, if that was all the bullet was saying, there was no need to say it at all):

 

Table 27-20—HE-SIG-A field of an HE MU PPDU, Bandwidth field

 

Set to 4 for preamble puncturing in 80 MHz, where in the preamble the only punctured subchannel is the secondary 20 MHz channel.

Set to 5 for preamble puncturing in 80 MHz, where in the preamble the only punctured subchannel is one of the two 20 MHz subchannels in the secondary 40 MHz channel.

Set to 6 for preamble puncturing in 160 MHz or 80+80 MHz, where in the preamble the only punctured subchannels are the secondary 20 MHz channel and zero to three of the 20 MHz subchannels in the secondary 80 MHz channel.  No more than two adjacent 20 MHz subchannels are punctured across 160 MHz or 80+80 MHz.

Set to 7 for preamble puncturing in 160 MHz or 80+80 MHz, where in the preamble the only punctured subchannels are one or both of the 20 MHz subchannels in the secondary 40 MHz channel and zero to three of the 20 MHz subchannels in the secondary 80 MHz channel.  No more than two adjacent 20 MHz subchannels are punctured across 160 MHz or 80+80 MHz.

 

Option B: ignore SP4 and instead allow >2 adjacent 20M subchannels to be punctured (note: I think this is closest to the direction in D6.0, since it has no such adjacency restriction)

 

Table 27-20—HE-SIG-A field of an HE MU PPDU, Bandwidth field

 

Set to 4 for preamble puncturing in 80 MHz, where in the preamble the only punctured subchannel is the secondary 20 MHz channel.

Set to 5 for preamble puncturing in 80 MHz, where in the preamble the only punctured subchannel is one of the two 20 MHz subchannels in the secondary 40 MHz channel.

Set to 6 for preamble puncturing in 160 MHz or 80+80 MHz, where in the preamble the only punctured subchannels are the secondary 20 MHz channel and zero to three of the 20 MHz subchannels in the secondary 80 MHz channel.

Set to 7 for preamble puncturing in 160 MHz or 80+80 MHz, where in the preamble the only punctured subchannels are one or both of the 20 MHz subchannels in the secondary 40 MHz channel and zero to three of the 20 MHz subchannels in the secondary 80 MHz channel.

 

Option C: ignore SP3 and SP4 and instead allow P80 not to be punctured at all and allow >2 adjacent channels to be punctured (note: I think this is Ross’s preference; it does however mean receivers are potentially falsely directed to not attempt to receive a content channel on some 20 MHz subchannels)

 

Table 27-20—HE-SIG-A field of an HE MU PPDU, Bandwidth field

 

Set to 4 for preamble puncturing in 80 MHz, where in the preamble the only punctured subchannel is the secondary 20 MHz channel.

Set to 5 for preamble puncturing in 80 MHz, where in the preamble the only punctured subchannel is one of the two 20 MHz subchannels in the secondary 40 MHz channel.

Set to 6 for preamble puncturing in 160 MHz or 80+80 MHz, where in the preamble the only punctured subchannels are the secondary 20 MHz channel and zero to three of the 20 MHz subchannels in the secondary 80 MHz channel.

Set to 7 for preamble puncturing in 160 MHz or 80+80 MHz, where in the preamble the only punctured subchannels are zero, one or both of the 20 MHz subchannels in the secondary 40 MHz channel and zero to three of the 20 MHz subchannels in the secondary 80 MHz channel (but at least one if no 20 MHz subchannel was punctured in the secondary 40 MHz channel).

 

Option D: same as option A1, except that allow puncturing in S80 but not P80 by extending the definition of 3 (note: this is a way to address Ross’s concern about being forced to puncture a 20M in P80 if actually all you want to do is puncture in S80; it does however potentially cause problems with implementations that assume they can process content channels in S80 in “case 3”)

 

Table 27-20—HE-SIG-A field of an HE MU PPDU, Bandwidth field

 

Set to 3 for 160 MHz and 80+80 MHz non-preamble puncturing mode, or for preamble puncturing in 160 MHz or 80+80 MHz, where in the preamble the only punctured channels are one or two of the 20 MHz subchannels in the secondary 80 MHz channel.  If two of the 20 MHz subchannels in the secondary 80 MHz channel are punctured, these are either the lower two or the higher two.

Set to 4 for preamble puncturing in 80 MHz, where in the preamble the only punctured subchannel is the secondary 20 MHz channel.

Set to 5 for preamble puncturing in 80 MHz, where in the preamble the only punctured subchannel is one of the two 20 MHz subchannels in the secondary 40 MHz channel.

Set to 6 for preamble puncturing in 160 MHz or 80+80 MHz, where in the preamble the only punctured subchannels are the secondary 20 MHz channel and zero to two of the 20 MHz subchannels in the secondary 80 MHz channel.  If two of the 20 MHz subchannels in the secondary 80 MHz channel are punctured, these are either the lower two or the higher two.  No more than two adjacent 20 MHz subchannels are punctured across 160 MHz.

Set to 7 for preamble puncturing in 160 MHz or 80+80 MHz, where in the preamble the only punctured subchannels are one or both of the 20 MHz subchannels in the secondary 40 MHz channel and zero to two of the 20 MHz subchannels in the secondary 80 MHz channel.  If two of the 20 MHz subchannels in the secondary 80 MHz channel are punctured, these are either the lower two or the higher two.  No more than two adjacent 20 MHz subchannels are punctured across 160 MHz.

 

Thanks,

 

Mark

 

--

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

Samsung Cambridge Solution Centre       Tel: +44 1223  434600

Innovation Park, Cambridge CB4 0DS      Fax: +44 1223  434601

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

 

On Thu, 9 Apr 2020 at 18:41, Mark RISON <m.rison@xxxxxxxxxxx> wrote:

This is the follow-up to the teleconf discussion just now, to

attempt to reach consensus on which puncturing patterns are and

are not allowed.

 

STRAWMAN PROPOSAL:

 

If the primary 20 MHz channel is the leftmost one, and x means

punctured, - means not punctured and ? means can be punctured,

then these correspond to only allowing the following patterns

in each case (obviously the patterns will get rearranged if the

P20 is not the leftmost one) for the value of the Bandwidth

field in the HE-SIG-A field of an HE MU PPDU:

 

4) -x--

5) --x- or ---x

6) -x--???? but not -x--xxxx

7) --x-???? but not --x-xxxx, or

   ---x???? but not ---xxxxx, or

   --xx???? but not --xxxxxx

 

During the teleconf I think I heard some variants proposed (calling

the one above z):

 

a) Allow full puncturing of S80, i.e.:

 

4) -x--

5) --x- or ---x

6) -x--???? including -x--xxxx

7) --x-???? including --x-xxxx, or

   ---x???? including ---xxxxx, or

   --xx???? including --xxxxxx

 

b) Don't allow any puncturing of S80 (maybe the D1.0 wording?), i.e.:

 

4) -x--

5) --x- or ---x

6) -x------

7) --x----- or

   ---x---- or

   --xx----

 

c) Allow "any 2" puncturing for the 7 case (I think I heard something

like that?), i.e.:

 

4) -x--

5) --x- or ---x

6) -x--???? [all x for ? TBD]

7) --x-???? where exactly one of the ?s is a x, or

   ---x???? where exactly one of the ?s is a x, or

   ----???? where exactly two of the ?s are an x, or

   --xx----

 

d) Or was it "at least 2" puncturing for the 7 case, i.e.:

 

4) -x--

5) --x- or ---x

6) -x--???? [all x for ? TBD]

7) --x-???? where between 1 and 3 (4 TBD) of the ?s are xs, or

   ---x???? where between 1 and 3 (4 TBD) of the ?s are xs, or

   ----???? where between 2 and 3 (4 TBD) of the ?s are xs, or

   --xx???? where between 0 and 3 (4 TBD) of the ?s are xs

 

Plus a variant to any of the above:

 

e) For any/all of the above, don't allow more than 2 adjacent xs

for the 160M case (obviously N/A for 80 and I think also for 80+80)

 

Please select your option (z or a-d, plus e) or otherwise describe

your understanding.

 

Happy Easter, and best wishes to you, your colleagues and your families,

 

Mark


--

Mark RISON, Standards Architect, WLAN   English/Esperanto/Français
Samsung Cambridge Solution Centre       Tel: +44 1223  434600
Innovation Park, Cambridge CB4 0DS      Fax: +44 1223  434601
ROYAUME UNI                             WWW: http://www.samsung.com/uk



--

Mark RISON, Standards Architect, WLAN   English/Esperanto/Français
Samsung Cambridge Solution Centre       Tel: +44 1223  434600
Innovation Park, Cambridge CB4 0DS      Fax: +44 1223  434601
ROYAUME UNI                             WWW: http://www.samsung.com/uk


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


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



--

Mark RISON, Standards Architect, WLAN   English/Esperanto/Français
Samsung Cambridge Solution Centre       Tel: +44 1223  434600
Innovation Park, Cambridge CB4 0DS      Fax: +44 1223  434601
ROYAUME UNI                             WWW: http://www.samsung.com/uk


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