Jack

I will give in to your challenge, and focus on outdoors setups :) But first,
I agree with you 100% about your comments; some venders saying "add more APs
will help" as well as " their smart systems will solve all their performance
issues.

We have been watching, talking and helping some of the people involved in
the so called "city wide wifi" and many if not all of them are simply not
adhering to the strengths of wifi and trying to make the protocol do things
it will not.  When we get calls like this the most common misconceptions is
they expect laptops to work miles away from the AP, Non-LOS, and Omni
installs to keep it simple. Then we put on our training hats to bring them
down to earth on what to really expect and how to design a system correctly.


Ok now to a design ideas, first I want to point out that each design is
going to be different but if the basic concepts are followed it will work
each and every time.
        1. Before any design can be started you really want to get a base
line as to what one is starting with. We always recommend rents hand held SA
and do field tests to see what other frequencies are out there being used
and do your best to triangulate and find out where they are coming from.
This data should be recorded on a map for future reference.
        2. 90% of our customers try to find the highest place in the middle
of the area they are trying to serve, while this could be the best choice
for a design it's not the only or best one. See if you can service the same
area from the outside in, for example if you are able to find three
locations on the outer rim of an area and use 120 degree antennas this will
increase your true coverage area exponentially as the installers now have
three locations to try vs. one. The latest customer we helped with a design
went from 25% install rate to 90%, thus it's well worth the costs for the
extra two locations up front where the costs of a truck role is about $50
minimum if the customer is installed or not! The first month this design
saved the customer well over $5000 in install fees plus he now over 100
customers generating income that he would not have at this point!
        3. Once the locations are found on the tower or building you REALY
want to avoid installing antennas anywhere near each other. On a tower try
to install the antennas with 10' vertical separation at a minimum, more is
better and vertical is more important than horizontal.  Also try to have
2-3' horizontal separation.  Ideally you want to create a spiral stair case
effect with the antennas. 
        4. This is where the setup is very important, you NEED to limit the
distance of each antennas to be practical. For example if you are in a rural
area this could be 10+ miles where you do not see any other radio signals
via the antenna, yes you want to test with an SA directly on the antennas to
confirm! In other locations this could be less than 1 mile.  As an example
let's assume a WISP wants to reach a max of 5 miles, having a person on the
tower to adjust the down tilt, you want to have an install at the 5 mile
mark (and center of antenna be is 180, 120 etc) and tune the strongest
signal at this point.  With this setup you can rest assure that the antennas
will only pick up interference from 5 miles in and not from 5 miles out. The
name of the game with wireless is single to noise levels, lots of signal
will do nothing if there is also lots of noise.
        5. The other major issue we see is self interference. If you have
more than one antenna at a locations, you want to test with the SA each
antenna how much it's sees from the other antennas. So for example if you
have three radios at a sight plug the SA in to one antennas and turn on one
radio then the other and see how much signal the antennas pick up from the
other radios and do the same for the other two.  The issue here is even
though each antennas/radio is on a different channel too much power from the
side radios can cause major problems with the receivers, basically
overloading them. This is the major reason you need to have good separation.

There are a few other tricks but I think I have hit on the major ones, I am
sure Jack will fill in if I missed something major. Anyway hope this helps
and if anyone wants to talk more you can hit me off list or call me or my
techs at the office.

Sincerely, Tony Morella
Demarc Technology Group, A Wireless Solution Provider
Office: 207-667-7583 Fax: 207-433-1008
http://www.demarctech.com 
 
This communication constitutes an electronic communication within the
meaning of the Electronic Communications Privacy Act, 18 USC 2510, and its
disclosure is strictly limited to the recipient intended by the sender of
this message. This communication may contain  confidential and privileged
material for the sole use of the intended recipient and receipt by anyone
other than the intended recipient does not constitute a loss of the
confidential or privileged nature of the communication. Any review or
distribution by others is strictly prohibited. If you are not the intended
recipient please contact the sender by return electronic mail and delete all
copies of this communication


-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf
Of Jack Unger
Sent: Thursday, February 14, 2008 2:17 AM
To: WISPA General List
Cc: [EMAIL PROTECTED]
Subject: Re: [WISPA Members] [WISPA] WLAN stress test uncovers
802.11performance problems

Yep. Very few people (less than 5%) who deploy dense wireless LANs
understand that the interference radius around each access point Omni
antenna is 4 times to 8 times (or more) GREATER than the usable  
communications radius. People end up putting too many access points too
close together resulting in a huge DECREASE in aggregate throughput and
REDUCED network reliability. These symptoms will only emerge once  traffic
levels build up and wireless frames are colliding with each other all over
the place. Then the network manager starts adding access points and makes
the problem worse.

On top of that, most wireless LAN vendors push the idea that throughput and
reliability problems will be SOLVED by adding more access points. The
wireless LAN vendors also tell the customer that their "smart" 
controller will help reduce the self-interference between the access points
by "intelligently" managing access point power and channel. Those of you who
don't live in a major city will have to use your imagination to follow this
example but this vendor argument is like saying that placing a traffic cop
with a whistle in the middle of a freeway that is moving at 2 miles per hour
will help to "manage" the flow of traffic and speed the traffic up; it
won't. The cop will only manage to distract the 
drivers and slow the traffic down even more.

So if you are deploying a wireless LAN that needs a lot of access points to
handle a many simultaneous users and high levels of user traffic and
different types of user traffic (like voice and data), what should you do to
get your WLAN to work and to meet management and end-user expectations?

Let's hear a few opinions from people about how to get good WLAN
performance. BTW, these very same principles apply outdoors in WISP
deployments.

After we hear from five or six people, I'll offer my solutions (if you all
don't beat me to it). ;-)

Oh yeah, one more thing... PLEASE post your solutions over on the WISPA
Members list. This is the kind of info that I feel paying WISPA members
deserve to have access to. In my view (I know this is 
"old-school" thinking these days but remember, I'm old), there really is no
such thing as a "free lunch". Those folks who support WISPA deserve to
receive information that is a cut above the "free information" (and
dis-information and advertising hype) that has distorted people's sense 
of reality these days.

Your normally quiet and humble servant,

jack





--------------------------------------------------------------------------------
WISPA Wants You! Join today!
http://signup.wispa.org/
--------------------------------------------------------------------------------
 
WISPA Wireless List: wireless@wispa.org

Subscribe/Unsubscribe:
http://lists.wispa.org/mailman/listinfo/wireless

Archives: http://lists.wispa.org/pipermail/wireless/

Reply via email to