[Server-devel] Technical questions

[James Cameron]

On Wed, Jan 20, 2010 at 06:10:34PM -0800, akleider at sonic.net wrote:
> I for one would be interested in knowing if you can comment about 
> specific hardware (hopefully things currently available) that might serve
> as examples of some of the things you've discussed:
> i.e.: access points, improved antenna systems, larger omni directional
> antennas

No, not really.  It's not a problem domain that can be trivially
addressed by specific hardware.  Success depends on many factors, not
just the hardware.  Also the hardware available to me in Australia is no
doubt quite different to the hardware available to others.

There are rules of thumb, and there is engineering.

A radiofrequency engineer (I'm an amateur radio operator) ... can
determine by theory and measurement the link budget [1] between an
access point and a laptop in varying positions, and then assess the
performance of a system as a whole; access points and laptops.

There are two main problems: received power too low, or too much noise.

If the received power is too low for the sensitivity of the receiver,
things can then be done to correct it; such as increasing sensitivity of
the laptop radio, changing position of laptop, orientation of laptop
antennas, obstacles in line of sight, obstacles in fresnel zone [2],
orientation of access point antennas, position of access point antennas,
and increasing the transmitted power.

a.  increasing sensitivity of the laptop radio; is not an option for you
unless you are changing which model of laptop is being used,

b.  changing position of laptop; is very obvious, and kids know how to
do that quite quickly, ... in some places it doesn't work, so they avoid
those places,

c.  orientation of laptop antennas; usually these should be vertical to
the ground, parallel with the access point antennas,

d.  obstacles in line of sight; if the access point antenna can be seen
by eyesight then this is a good start,

e.  obstacles in fresnel zone; usually only relevant for long distances,
see the theory article, ... it is why windows don't always work for
signal propagation,

f.  orientation of access point antennas; again normally these should be
vertical,

g.  position of access point antennas; the higher they are, the greater
coverage (and intereference) they will have, ... on the other hand the
longer the cable is between the access point and the antenna, the more
loss is felt,

h.  increasing the transmitter power; within the constraints of the
country regulations and the limits of the access point.

There are environmental considerations that also affect a link budget
... extreme humidity or rainfall can increase signal loss, direct
sunlight can increase noise, and other radio systems outside the site
can increase noise.

On the other hand, if received power is adequate and the problem being
experienced is noise from other radios in the other access points or
laptops, then some of the above can be *reversed* to reduce the problem.
I think I mentioned that in detail in my previous mail.

There's no simple answer, sorry.

But if your school looks very much like another school, and their
hardware choices worked well, then chances are the same choices will
give a similar result. 

References:

1.  http://en.wikipedia.org/wiki/Link_Budget ... "the accounting of all
of the gains and losses from the transmitter, through the medium (free
space, cable, waveguide, fiber, etc.) to the receiver in a
telecommunication system"

2.  http://en.wikipedia.org/wiki/Fresnel_zone

-- 
James Cameron
http://quozl.linux.org.au/