Two access point a/b/g in bridge mode

Hi everyone,
we need a very reliable wireless connection between a laptop and a
small network of two other computers. We first tried with an ad-hoc
network with wavelan cards, but since usually there are walls and a lot
of interferences between the two ends, the communication is very poor.

We switched to one ap to wavelan pcmcia card, but this still has
problems, so we decided to use two access points in bridge mode or
server-client mode (which one is better?), but it is difficult to find
two access points with those characteristics (and that we can find in
Italy!)...

- 802.11a and 802.11g mode
- bridge and/or client-server modes
- one must have a switch, because it have to be connected to two pc

any ideas?

On 31 May 2005 02:59:20 -0700, "MadMage" <(E-Mail Removed)> wrote:

>Hi everyone,
>we need a very reliable wireless connection between a laptop and a
>small network of two other computers. We first tried with an ad-hoc
>network with wavelan cards, but since usually there are walls and a lot
>of interferences between the two ends, the communication is very poor.

Wavelan PCMCIA cards are the original names for the
Orinioco/Lucent/Avaya/Agere/Proxim/Gold/Silver/Whatever cards. The
really really early ones will only do 1 and 2 Mbits/sec. Later units
has firmware and driver issues and were not tested with later
operating systems. They usually function, but I'm not sure I would
consider them the best choice if you insist on a reliable connection.

In my never humble opinion, you are NOT going to get any form of
reliable wireless communications going through multiple walls and in
the presence of interference. 802.11 was never designed to be jam
proof in the presense of reflections with weak signals.

You haven't mentioned performance (speed), but if that's also a
requirement, I consider it impossible. You can improve reliability
somewhat at the cost of speed. I suggest you:
Set the maximum speed to 2 Mbits/sec
Turn on CTS/RTS flow control in the access point
Decrease the fragmentation threshold to about 500 bytes.
These changes will increase the sensitivity, do a better job dealing
with interference, and decrease the size of the packets so that they
have a better chance of surviving in the presence of intereference.
However, it will be very slow on thruput.

>We switched to one ap to wavelan pcmcia card, but this still has
>problems, so we decided to use two access points in bridge mode or
>server-client mode (which one is better?), but it is difficult to find
>two access points with those characteristics (and that we can find in
>Italy!)...

What you are looking for is a "transparent bridge" or "workgroup
bridge". These are wireless bridge devices that will pass more than
one MAC address. For example, WAP54G, DWL900AP+, etc. Some of the
"game adapters" (WET11) will also work with more than one MAC address.

>- 802.11a and 802.11g mode

If you really have interference problems, moving to 5.6GHz might not
be a bad idea.

>- bridge and/or client-server modes

There is no client server mode. There are:
Anfrastructure
Ad-hoc
WDS (wireless distribution service)
Client mode with one MAC address
Transparent bridge
Client mode with more than one MAC address (same as
transparent bridge mode.

>- one must have a switch, because it have to be connected to two pc

Most wireless transparent bridges have only a single ethernet
connection because they assume the customer will supply the switch.

>any ideas?

If you forget about 802.11a 5.7GHz, I suggest a Linksys WRT54G with
Sveasoft Alchemy firmware. This will give you *ALL* the aformentioned
modes to experiment. See:
http://www.sveasoft.com/content/view/3/1/
However, I don't have a solution to the intereference or wall
penetration problems. Those will be there no matter what hardware you
select.


--
Jeff Liebermann (E-Mail Removed)
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 AE6KS 831-336-2558

On Tue, 31 May 2005 09:15:05 -0700, Jeff Liebermann
<(E-Mail Removed)> wrote:

>There is no client server mode. There are:
> Anfrastructure

That should be Infrastruture mode.

(That's what happens if I forget to run the spellin chequer).


--
Jeff Liebermann (E-Mail Removed)
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 AE6KS 831-336-2558

"Jeff Liebermann" <(E-Mail Removed)> wrote in message
news:(E-Mail Removed)...
> On 31 May 2005 02:59:20 -0700, "MadMage" <(E-Mail Removed)> wrote:

<snip>

> You haven't mentioned performance (speed), but if that's also a
> requirement, I consider it impossible. You can improve reliability
> somewhat at the cost of speed. I suggest you:
> Set the maximum speed to 2 Mbits/sec
> Turn on CTS/RTS flow control in the access point
> Decrease the fragmentation threshold to about 500 bytes.
> These changes will increase the sensitivity, do a better job dealing
> with interference, and decrease the size of the packets so that they
> have a better chance of surviving in the presence of intereference.
> However, it will be very slow on thruput.

<snip>

Hm, interesting.
For regular usage, what would you recommend the RTS, CTS and Frag settings
be? Defaults good most of the time?
I know it's not much to go on and site will vary but, let's say for a
regular small apartment line of site and 2nd situation would be a 3 story
wood house?
Just curious if anyone tunes/tweeks with these settings,
experience-wise...... ;-)

Scotty

On Wed, 01 Jun 2005 04:33:55 GMT, "Scott Nelson - Wash DC"
<(E-Mail Removed)> wrote:

>
>"Jeff Liebermann" <(E-Mail Removed)> wrote in message
>news:(E-Mail Removed).. .
>> On 31 May 2005 02:59:20 -0700, "MadMage" <(E-Mail Removed)> wrote:
>
><snip>
>
>> You haven't mentioned performance (speed), but if that's also a
>> requirement, I consider it impossible. You can improve reliability
>> somewhat at the cost of speed. I suggest you:
>> Set the maximum speed to 2 Mbits/sec
>> Turn on CTS/RTS flow control in the access point
>> Decrease the fragmentation threshold to about 500 bytes.
>> These changes will increase the sensitivity, do a better job dealing
>> with interference, and decrease the size of the packets so that they
>> have a better chance of surviving in the presence of intereference.
>> However, it will be very slow on thruput.

>Hm, interesting.
>For regular usage, what would you recommend the RTS, CTS and Frag settings
>be? Defaults good most of the time?

The default settings for all wireless routers is to turn *OFF* CTS/RTS
flow and fragmentation threshold. CTS/RTS flow should be enabled only
if there are a large number of users battling for the attention of an
access point. It also helps if there is interference. Fragmentation
helps by creating smaller packets, which have a higher probability of
arriving in the presence of interference. I don't really have a good
set of values in you want to enable these. The smallest packet is
256bytes, but that's probably too small. 512bytes will work well
enough. However, don't expect playing with these to improve the
performance and speed.

>I know it's not much to go on and site will vary but, let's say for a
>regular small apartment line of site and 2nd situation would be a 3 story
>wood house?

Neither justifies turning on either CTS/RTS or fragmentation, unless
you have an army of users or tons of interference. Leave the values
at the defaults, which are essentially "off".

>Just curious if anyone tunes/tweeks with these settings,
>experience-wise...... ;-)

I do for WISP (wireless ISP) service. Anything installed outdoors, in
a crowded city, will pickup interference. If the packet loss appears
to be climbing, I start sending smaller packets. If I have a large
number of simultaneous users (over 30), then CTS/RTS flow is required.

Incidentally, the value for CTS/RTS flow is the minimum packet size
that will *NOT* use CTS/RTS flow. It makes no sense to initiate flow
control negotiations for sending extremely small packets. One can
resend 3 or 4 small packets in the same time that it takes to
negotiate one flow control exchange. So, the setting determins the
threshold of where larger packets get flow control and smaller packets
simply get sent, and resent if they fail.

>Scotty


--
# Jeff Liebermann 150 Felker St #D Santa Cruz CA 95060
# 831.336.2558 voice http://www.LearnByDestroying.com
# (E-Mail Removed)
# (E-Mail Removed) AE6KS