Maximum connection speed over wireless

[ve9mdb]

Hi,

This question might already be asked but I did not found it with the search function.

My question : What is the maximum possible speed for data or voice over a wireless connection, assuming that both points are at about -55 dBm ?
I am currently using some Teletronics products ( TT5800 ) and they state that we can establish a 108 Mbps connection while using turbo mode ( Atheros function making the radio to use 2 differents channels to establish the connection).
Is that right or not totally right ?
If we were using two NS5 to do the same thing ... what would be the improvement ?

Thanks for your help.

[jp498]

108mbps connection is fairly meaningless. If a 54mbps connection tops out at 20-something mbit, then you'll get 40-50mbit out of it in ideal conditions.

These tests are typically done with full 1500byte packets too. If you have lots of voip or p2p doing 100 byte packets fully utilizing the radio, then you'll have correspondingly less performance which can be depressing if you're expecting too much out of it. You don't have than problem on a wired network.

If you want something of high speed, get something that can combine packets. Alvarion's VL was among the first things to do this. I can get a good 30+mbits of general internet traffic over a ptp VL link, which is all different packet sizes over a 54mbps radio modulation. Trango can do something similar with their backhauls but it's details are not part of the user interface. Mikrotik Nstreme does something similar with it's framing policy options. If you did Mikrotik, most of their cheaper boards can't handle 50mbit; you have to get an AH board or better or use a PC. You also need a good steady signal like you suggest. Big packets have more information to potentially get messed up by interference, so steady signal and clean spectrum are important.

[ve9mdb]

Right, ok now I better understand how speed work.
Basically, it is the same as an hard drive which have only x cluster of xxx bytes and even if the data take only x byte it will take the whole cluster. So that being said, this would mean that the connection speed is still 54 Mbps or 108 Mbps but if the paquet being sent from one point to another is not "full", the real data speed will be lower.

(I'd try my best explaining in english as I am french ... but I think the concept is being well explained isn't it?).

This bring me to ask : what would happen in the case of streaming (like VOIP) ? I do use Vonage and plan to use it over our wireless link. The first 2-3 weeks of testing went well but I was almost alone on the link. We now have a major issue with our equipement because I can't get more than 80 kb/sec out of the link ... I think there is something wrong with our radio so I'll be switching them and put an amplifier as well .... hopefully everything come back to what we had at first.

Any comments about VOIP ?

thanks

[WHT]

put an amplifier as well .... hopefully everything come back to what we had at first.

Any comments about VOIP ?

thanksAmplifiers are for WISP wannabes...they create more problems than they solve.

VoIP is often problematic on wireless. It takes a lot of tweaking to get it to work, QoS and all at the head-end.

[ve9mdb]

Thanks for the reply WHT.

Up to date, we haven't had any problems with Vonage ...and I wish it stay as good as it was last month (before the faillure). Would you recommend a specific product for wireless VOIP applications ?

Speaking of amplifiers : I like them for the "receive amplification" it give. You may get up to 20 dB of gain on receive. Thus mean you can raise a signal over the noise level just enough to establish the connection ( as long the other end have a good reception of the AP).

I was using Rflinx amplifiers and they are well made. Haven't found troubles yet with them. I think they have a clean rf and are not overmodulate.
Also, good radios with enough power are costly (in exception of Ubiquiti). This is why I am interested in Ubiquiti systems and want to try them out. We will be ordering a bunch of Picostation and Nanostation soon ( mostly 5 Ghz and some 2.4 Ghz ).

[WHT]

Speaking of amplifiers : I like them for the "receive amplification" it give. You may get up to 20 dB of gain on receive. Thus mean you can raise a signal over the noise level just enough to establish the connection.

The biggest problem is they also raise the noise level. So the difference of the SNR remains the same.

Here in the U.S., you won't find many, if any at all. transmitters certified for an external amplifier.

[ve9mdb]

Forgot that one ... about the noise floor level that is being raised by amplifier !

Thanks WHT.

Does Ubiquiti produce any backhaul radios ???

[dacost4]

Thank :)

[WHT]

Does Ubiquiti produce any backhaul radios ???
Any UNBT radio can be used for a backhaul, but alot of people use the PS5 that is certified for a 23 dBi antenna.

On the other hand, the Bullet 5 and HP is certified for a 30 dBi antenna. The only downside is its lower temperature range. Sticking it inside a PVC tube with some insulation might fix that...or even a small wire wound 5 watt power resistor connected to a thermostat stuffed inside the tube as well.

[Stenn]

What's the maximum distance at which the equipment will keep us at least 54mbs

[WHT]

Off the top of my head, a pair of Powerstations are good for about 3 miles.

Keep in mind a 54 Mbps data speed is only 22 Mbps throughput by the time you subtract the wireless protocol overhead.

[remuda]

Any comments about VOIP ?

thanks

I'm using voip over six bullet5's w/o QoS to my asterisk server. No problems with the exception of echo (until the algos kick in).