Hi,
Yes, we've adopted it in a scenario where we share a building with
another institution. They use 2 channels and we use the other 2........it
works for us.
I think you'll find that it's a fairly common type of deployment.
..............thx...............J
James Savage York University
Senior Communications Tech. 108 Steacie Building
[EMAIL PROTECTED] 4700 Keele Street
ph: 416-736-2100 ext. 22605 Toronto, Ontario
fax: 416-736-5701 M3J 1P3, CANADA
Lelio Fulgenzi <[EMAIL PROTECTED]>
01/01/2007 09:58 PM
Please respond to
"802.11 wireless issues listserv" <WIRELESS-LAN@LISTSERV.EDUCAUSE.EDU>
To
WIRELESS-LAN@LISTSERV.EDUCAUSE.EDU
cc
Subject
[WIRELESS-LAN] using four channels instead of three
I ran across this article a while back (I'm pretty sure it was a result of
searching the list archives) and I'm wondering if anyone out there has
adopted this method in practice.
http://findarticles.com/p/articles/mi_zdext/is_200211/ai_ziff33684
4 Simultaneous Channels Okay For 802.11b
ExtremeTech, November, 2002 by Jim Louderback
Note: Last summer, before starting here at Ziff Davis Media, I did some
work with a small 802.11b startup called Cirond. The following story is
based on a white-paper analysis done by CTO Mitch Burton. After I read it,
I wanted to tell the world. I'll make no claims on whether Cirond's
forthcoming products are useful, but its Channel Overlap analysis is
exciting. But in the interest of full disclosure, note that they did pay
me for some of the work I did ? although that was almost six months ago.
So take this story with the proverbial grain of salt. I find it exciting,
you may not.
Current thinking in the 802.11b wireless space is that only three of the
11 channels used by wireless hubs in the US can be allocated
simultaneously. But that's wrong, according to Mitch Burton, CTO of Cirond
Networks. In fact, 4 of the 11 channels in North America, and 5 of the 13
in Europe can be safely used without significant interference or crosstalk
? and this has significant ramifications for multi-access point
deployments.
In North America, the 802.11b spectrum ranges from 2400MHz to 2483MHz, and
is divided up into 11 channels from 2412MHz to 2462MHz, spaced 5MHz apart.
Thus However, each channel is 22MHz wide, so as you can imagine, there is
great overlap. Channel 1, for instance, is centered at 2412MHz, but
extends out from 2401MHz to 2433MHz. Channel 6 is centered at 2437MHz,
extending from 2426MHz to 2448MHz.
In a multi-access point installation, where overlapping channels can cause
interference, dead-spots and other problems, Channels 1, 6 and 11 are
generally regarded as the only safe channels to use. Since there are 5
5MHz channels between 1 and 6, and between 6 and 11, or 25MHz of total
bandwidth, that leaves three MHz of buffer zone between channels.
Note that wireless access points generally radiate waves in a sphere
around the access point, attenuated by walls, cubicle material, ceilings
and floors. With just three channels to work with, it can become difficult
to deploy wireless access around a single or multi-floor location while
only reusing those 3 frequencies
This picture shows how you would have to overlap Access Points with only
three channels on a single floor:
It's possible, but when you add in a third dimension, it starts to not
work out:
You cannot have two access points per floor in this scheme, which makes
adding wireless into a multi-floor building a very tricky prospect.
But what would happen if you could have four non-overlapping channels
instead of three?
Your flat access point distribution pattern would look like this:
And your 3D distribution across different floors gets much easier:
Now each floor can have two access points within the same coverage area,
and not interfere with each other. This is a much easier set up to design,
plan and maintain.
But according to conventional wisdom, a four channel model involves
overlap and interference, per the chart below:
Channel Start Fqy Mid Fqy End Fqy 1 2401 2412 2423 4 2416 2427 2438 8 2436
2447 2458 11 2451 2462 2473
Channel 1 ends at 2423MHz, while channel 4 starts at 1416MHz, and that's a
significant overlap. Channel 4, for example, only has 13MHz of its 22MHz
unimpeded by overlapping frequency.
But a standard mathematical analysis misses two fundamental aspects of
wireless frequency use patterns of 802.11b equipment. First, an entire
22MHz is not simply swallowed up in a rectangular pattern with power on
the vertical axis and frequency on the horizontal ? instead it's more of a
parabola, centered around the midpoint of the frequency. Thus, as you get
further away from the center, the power drops off substantially.
According to Burton's analysis, when three channels separate 802.11b
frequencies, there is only about 4% of interference. This is the case
between frequencies 1 and 4, and 8 and 11. Between 4 and 8, the
interference is substantially less than 1%.
And today's wireless equipment uses filters to further reduce interference
issues.
Essentially, according to Burton's analysis, you can safely use four
channels in North America, and five channels in Europe, instead of three
and four, when planning your multi-access point network. The advanced
filtering of today's 802.11b equipment, coupled with the power falloff
seen in the parabolic curves exhibited at the outer ranges of a 22MHz wide
channel means that you shouldn't run into problems.
The white paper provides additional details on signal to noise
calculations, additional mathematics and other details. But fundamentally,
what we've known to date is wrong. It looks like four channels with some
overlap will solve a lot of our 802.11b deployment problems.
********** Participation and subscription information for this EDUCAUSE
Constituent Group discussion list can be found at
http://www.educause.edu/groups/.
**********
Participation and subscription information for this EDUCAUSE Constituent Group
discussion list can be found at http://www.educause.edu/groups/.
