"More channels means more capacity" is not true.  Because the number of
null subcarriers is fixed and independent of channel width, wider channels
will make more efficient use of the spectrum.  You'll get the most
capacity out of the 802.11ac spectrum by using (6) 80MHz channels and (1)
20MHz.  Of course, a variety of conditions and design choices affect
capacity, not just channel width.  That's why we don't build networks that
way.

It's no surprise that this contention was generated by a couple very
generalized questions.  The topic is way too complex to for a thorough
discussion in this format.  Any answer of reasonable length is going to
leave a host of assumptions unstated.  If yours are the same as mine we'll
probably agree.  If yours are different from mine you'll think I'm an
idiot.

-----Original Message-----
From: The EDUCAUSE Wireless Issues Constituent Group Listserv
[mailto:[email protected]] On Behalf Of Curtis K. Larsen
Sent: Tuesday, September 26, 2017 3:17 PM
To: [email protected]
Subject: Re: [WIRELESS-LAN] Two RF Questions

>From the Cisco/Apple Design Guide Here:  https://goo.gl/5bGWks

"It is therefore not yet recommended to use 80 MHz channel width design.
If necessary, it should only be considered for low AP density deployments
where co-channel interference can be easily avoided."

I personally like the approach here:  https://goo.gl/FcPHFq

- More channels means more capacity
- 80MHz - small deployment with no interference - 40MHz - with thick
walls, one floor, and/or small deployments - 20MHz - by default


Thanks,

Curtis

________________________________________
From: The EDUCAUSE Wireless Issues Constituent Group Listserv
<[email protected]> on behalf of Jeffrey D. Sessler
<[email protected]>
Sent: Tuesday, September 26, 2017 1:08 PM
To: [email protected]
Subject: Re: [WIRELESS-LAN] Two RF Questions

Jake,

GT's statement doesn't speak to the quality of the university's WiFi
design, only that this change made a difference. Again, without the
context, I still assert it's meaningless.

Jeff

From: "[email protected]"
<[email protected]> on behalf of Jake Snyder
<[email protected]>
Reply-To: "[email protected]"
<[email protected]>
Date: Tuesday, September 26, 2017 at 11:49 AM
To: "[email protected]"
<[email protected]>
Subject: Re: [WIRELESS-LAN] Two RF Questions

Jeff,
Take in context that GT works for a company that builds a tool to quantify
wireless problems based in depth packet analysis.  So when he says he sees
35% improvement, there's a lot of data that goes into it.
Sent from my iPhone

On Sep 26, 2017, at 12:41 PM, Jeffrey D. Sessler
<[email protected]<mailto:[email protected]>> wrote:
"After a switch to 20 MHz only, there was a 35% improvement in end-user
Wi-Fi experience."

I would argue that this is a meaningless statement without context, and
probably a bad question to ask a user in the first place. What does the
user think "experience" means i.e. the ability to connect or how well
their speedtest performs? It's not specific enough to draw a conclusion.

For example:

  1.  If 1/3 of my users had a device that could not associate because of
how the primary channel was selected in a 40 or 80 MHz wide deployment,
then those people would not be happy. If I then change to 20 MHz only,
allowing those users with the problematic device to connect, there will
obviously be a significant improvement in those user's WiFi experience.
The other users may still be happy because they can still connect.
  2.  If my buildings are open-concept (no walls/doors), and I have 24
AP's on a 1000 sq/ft floor plan, and statically set to 80 MHz channels,
then the end-user WiFi experience is going to be really poor. If I then
switch all those APs to 20 Mhz only, of course it's going to be a huge
improvement. Clearly, it was a poor design, and less about the channel
width and more about the person who thought they knew better.

Of course, if the survey questions were more specific, and had questions
like, "Do you consistently receive the highest 4K stream rate from
NetFlix", the satisfaction for this question may trend down.

Jeff



From:
"[email protected]<mailto:[email protected].
edu>"
<[email protected]<mailto:[email protected].
EDU>> on behalf of GT Hill <[email protected]<mailto:[email protected]>>
Reply-To:
"[email protected]<mailto:[email protected].
edu>"
<[email protected]<mailto:[email protected].
EDU>>
Date: Tuesday, September 26, 2017 at 8:47 AM
To:
"[email protected]<mailto:[email protected].
edu>"
<[email protected]<mailto:[email protected].
EDU>>
Subject: Re: [WIRELESS-LAN] Two RF Questions

I know that this is just one example, but I was at a large university site
(Cisco Wi-Fi) that was running 20/40 channelization. After a switch to 20
MHz only, there was a 35% improvement in end-user Wi-Fi experience.

Jake - One feature that I think many people agree is missing in FRA is the
ability to dynamically turn off a radio. In some cases an extra radio in
either band hurts more than it helps.

And to just stir the pot a bit, I wish there were SMALLER than 20 MHz
channelization. In many high density environments 20 MHz is just too big.
Give me some more radios at smaller channel sizes and I'll show you a
spectacular Wi-Fi network. :-)

GT

From: The EDUCAUSE Wireless Issues Constituent Group Listserv
<[email protected]<mailto:[email protected].
EDU>> on behalf of Jake Snyder
<[email protected]<mailto:[email protected]>>
Reply-To: The EDUCAUSE Wireless Issues Constituent Group Listserv
<[email protected]<mailto:[email protected].
EDU>>
Date: Tuesday, September 26, 2017 at 9:39 AM
To:
<[email protected]<mailto:[email protected].
EDU>>
Subject: Re: [WIRELESS-LAN] Two RF Questions

My challenge, as I've stated on this list before, is that Mac OS X
preferences width in its AP selection criteria.  So while you may get more
capacity, in a large Mac environment you lose most of that with Macs
hanging onto APs linger and having to rate-shift down to slower PHY speeds
due to that AP having a wider channel than its neighbors. Yes, it's dumb.
But he's the driver of that lambo.

Also, couple that with increasing the noise floor by 3db every time you
double the channel width and there are many cases where your lambo just
spins it's tires.  All that power and you can't hook it up.

Remember that spectrum is our constraining resource.

Figure out what width of channel you can run in a building, and run that.
That's the best use of spectrum and sure to give you the most smiles/hour
on your lambo.

I really like what cisco did with FRA.  Give me the ability to see what it
thinks the overlap is.  I would LOVE to see the same with DBS, and give me
what width it thinks all the APs in the building can pull off.

Sent from my iPhone

On Sep 26, 2017, at 8:19 AM, Jeffrey D. Sessler
<[email protected]<mailto:[email protected]>> wrote:
It's surprising to me that anyone would purchase a Lamborghini, then
disconnect ten of the twelve cylinders and drive it at 25 mph on the
autobahn.

When I see static 20 MHz channels, or using 40 MHz in only limited areas,
I wonder what's behind the purposeful neutering of the system. If you are
a Cisco customer running 8.1 or above, and not using DBS (Dynamic
Bandwidth Selection), then it's the equivalent of the Lamborghini above
running on only two cylinders.

Don't miss out on the significant advancements in bandwidth management.
Free those resources spent doing point-in-time simulation and surveys for
something the software doesn't already do far better at. I promise, DBS
won't hurt a bit and your users will thank you a hundred times over.

Jeff


From:
"[email protected]<mailto:[email protected].
edu>"
<[email protected]<mailto:[email protected].
EDU>> on behalf of "Street, Chad A"
<[email protected]<mailto:[email protected]>>
Reply-To:
"[email protected]<mailto:[email protected].
edu>"
<[email protected]<mailto:[email protected].
EDU>>
Date: Tuesday, September 26, 2017 at 6:59 AM
To:
"[email protected]<mailto:[email protected].
edu>"
<[email protected]<mailto:[email protected].
EDU>>
Subject: Re: [WIRELESS-LAN] Two RF Questions

What is your reasoning behind not wanting 40 megahertz channels if you
have plenty of overhead with your channel utilization?  People saying you
should or should not do something without Gathering any type of metric
worry me.

On Sep 25, 2017 3:28 PM, Chuck Enfield
<[email protected]<mailto:[email protected]>> wrote:

1.      Enable it in places to check for radar events.  If you get few,
then yes.  Client devices are almost fully capable now.  Hidden SSID's are
the only issue.  Some clients don't probe on DFS channels, and will only
respond to beacons.  Make sure 2.4 is usable for the small number of
incompatible devices.

2.      No.  Don't even consider 40MHz unless you're using almost all the
DFS channels, but even then you'll probably have to disable it in some
high density areas.



From: The EDUCAUSE Wireless Issues Constituent Group Listserv
[mailto:[email protected]] On Behalf Of David Blahut
Sent: Monday, September 25, 2017 3:17 PM
To:
[email protected]<mailto:[email protected]
DU>
Subject: [WIRELESS-LAN] Two RF Questions



Greetings,

I have two hopefully simple RF related questions:

1.  Should I enable the extended UNII-2 channels campus wide?

2.  Should I enable 40Mhz channel width campus wide?

In other words what are you doing on your campus and what is the "best
practice?



Our wireless infrastructure:



3 Cisco 5508s running 8.2.141.0



20 - 3800 APs

368 - 3700 APs

414 - 3600 APs

8 - 3500 APs

7 - 1810 APs

32 - 1142 APs



Prime 3.1.0



Thanks for your input.

David

********** Participation and subscription information for this EDUCAUSE
Constituent Group discussion list can be found at
http://www.educause.edu/discuss.
********** Participation and subscription information for this EDUCAUSE
Constituent Group discussion list can be found at
http://www.educause.edu/discuss.


________________________________

This e-mail message (including any attachments) is for the sole use of the
intended recipient(s) and may contain confidential and privileged
information. If the reader of this message is not the intended recipient,
you are hereby notified that any dissemination, distribution or copying of
this message (including any attachments) is strictly prohibited.

If you have received this message in error, please contact the sender by
reply e-mail message and destroy all copies of the original message
(including attachments).
********** Participation and subscription information for this EDUCAUSE
Constituent Group discussion list can be found at
http://www.educause.edu/discuss.
********** Participation and subscription information for this EDUCAUSE
Constituent Group discussion list can be found at
http://www.educause.edu/discuss.
********** Participation and subscription information for this EDUCAUSE
Constituent Group discussion list can be found at
http://www.educause.edu/discuss.
********** Participation and subscription information for this EDUCAUSE
Constituent Group discussion list can be found at
http://www.educause.edu/discuss.
********** Participation and subscription information for this EDUCAUSE
Constituent Group discussion list can be found at
http://www.educause.edu/discuss.
********** Participation and subscription information for this EDUCAUSE
Constituent Group discussion list can be found at
http://www.educause.edu/discuss.
********** Participation and subscription information for this EDUCAUSE
Constituent Group discussion list can be found at
http://www.educause.edu/discuss.

**********
Participation and subscription information for this EDUCAUSE Constituent
Group discussion list can be found at http://www.educause.edu/discuss.

**********
Participation and subscription information for this EDUCAUSE Constituent Group 
discussion list can be found at http://www.educause.edu/discuss.

Reply via email to