First of all this is my first post in the CWNP forum and I have to say that it is a great pleasure to read your posts, a lot of good info. I have been working with WLAN for about 1.5 years and I really love it. At this moment I'm studying for CWNA and I want to do the test within 2 months.
Anyway, let's go to my question.
In a co-located environment when you have 2 APs in the same channel in my understanding it works like 1 AP because the CSMA/CA prevents two devices to transmit on the same time. There is no RF interference.
Now let's suppose the same 2 APs but with adjacent channels (e.g 1 and 2 or 1 and 3), how it works in this scenario? Is carrier sense preventing interference between AP/STA or is there a chance to have a collision? Is NAV from AP/STA in one channel set with the duration value from AP/STA in a adjacent channel?
I hope you can understand my point, sorry for my poor English. :-//
Another question, in Brazilian's regulatory domain (ANATEL), channels 1-13 are allowed in 2.4GHz band (Same as EU). Do you have experience using channels 1, 5, 9 and 13 for APs deployment? Thinking about performance, is there a chance to have better results than channels 1, 6 and 11 using a typical VoWLAN cell coverage (20% overlap, -70dBm cell boundaries and 19dB channel separation)?
For virtually all practical purposes two co-located BSSs in consecutive 2.4 GHz channels (e.g. 1 and 2) interoperate the same way that two BSSs some distance from each other but in the same channel. In both cases frames from one BSS are receivable by the other BSS but are weaker. The receiver cannot tell if the weakness is due to a nearby transmitter with a 5 MHz channel center separation or a not so nearby transmitter with some meters of geographical (or BSS center) separation.
There is nothing in most frames to determine the channel center of the transmission. A few frames, such as beacons, carry the channel number in an information element for just this reason.
Channel separation of 1, 6, and 11 is said to be ideal regardless of the relative location of the three BSSs using those channels. Channel separation of 1, 5, 9, and 13 could work just as well but with more qualifiers, the primary one being relative location and the second one being radio power level.
Channel, power, distance -- it's all about frequency reuse.
I hope this helps. Thanks. /criss
I did a few deployments with the 4 channels scheme you describe, it seems to be more in fashion in Sweden than in any other country in Europe.
Works properly as long as you respect the basics of wireless that Chris mentions.
Nevertheless, I often found that many clients in a public area used default drivers, which are mostly set to use channels up to 11 and don't behave well with higher channels.
Thanks for your answers Criss and Jerome.
The use of Channel 13 is really an issue regarding clients config, most of them come configured for USA domain (channels 1-11). I had an situation where the latest drivers were installed in the customer cards but the manufacturer had wrong values for Brazilian country code, it was only "hearing" channels 1-11 and only HR-DSSS BSSs. My solution was to use German country code for the cards.
Most of the PDAs also have this problem and in some cases there is no fix.
I have just realized all this after the customer complained that their clients had low signal level even being close to the AP. In fact the clients were connected to the closest AP using channels 1, 5 or 9.
I am sure that most of the guys here working in "1-13 countries" have seen that. X-D
One more question.
Do you know which limits the vendors use for power to assume that the medium is busy? Is it stated in the .11 standard?
Is there such a value ? WHich would imply that we rely on some RSSI based reading ?
Or would we use SNR also, which would mean that, whatever the noise level is, we could transmit if we can be x dB over this level, and if we don't READ a frame being sent by another device ?...
I believe the standard calls for 802.11b to use one of three CCA (Clear Channel Assessment) modes, based on a combination of either energy detect alone or receiving a valid signal. If there is a valid signal, the required detection level varies from -76 dBm to -70 dBm depending on the transmitter power of the detecting device.
Based on energy detect alone for 802.11b, there may not be any required level.
For 802.11g, valid signal, -76 dBm.
When I talked to a Cisco radio engineer, I believe he gave me a different figure than -76 dBm for their AIR-AP1232AG APs, that was closer to -71 dBm.
In bridge mode (AIR-AP1232AG) , with at least one version of the firmware, you can set the CCA level, but not in AP mode.
Take these figures as approximate, since I'm traveling, and don't have my books, IEEE standards, or notes.
One more question.
Do you know which limits the vendors use for power to assume that the medium is busy? Is it stated in the .11 standard?[/quote]