Forum

Can someone please explain why I see these results below? I prepared a lab for my students trying to demonstrate the claim that a 802.11b device will drop the data rate for the network. The lab failed miserably.

Using a 802.11B only client I connected to multiple vendors access points with no noticeable decrease in data rate for the other 802.11G clients. People were operating at 54Mbps while only my B client was at 11Mbps.

It just doesn't see true that a B only client drops all the clients in a G network to 11Mbps. I understand it may have to do with RTS/CTS, but if someone can explain this to me I'd greatly appreciate it!

Criss_Hyde Escribi?3:

Hi Freedom900 of Stockholm:

Access points typically use a mix of transmission speeds fit for the destination address and purpose of each frame. This might explain the variety you are seeing.

Each station, whether client or access point, has its own list of inherently supported rates, built in and unconfigurable, and published "supported rates", theoretically configurable. A station will always receive (demodulate) frames at its inherently supported rates but will only transmit frames at its published supported rates. Each station records the supported rates list of each of its link partners and further restricts its transmissions to a rate supported by its intended destination. Lastly each station maintains for each of its link partners a history of which one of several available mutually supported rates have adequate performance and may adjust the chosen data rate for that partner up or down in response to a changing RF environment.

I hope this helps. Thanks. /criss

Wellllll... it's not the data rate that drops. It's the throughput. Your .11g clients will maintain the same data rate connection but, due to protection mode, their throughput will suffer when .11b clients associate to the AP the .11g clients are on. So, if you're just using the supplicant software, you'll never know you've got a .11b client in your .11g environment. You have to use a network throughput analysis tool to see that impact. There are many free tools out there that will show you this. Also, your supplicant software may have a "Site Survey" mode. If it does, you could run it and that would show the decrease in throughput.

Joel

An entire BSS dropping its data rate to accommodate an 11b device is a well propagated myth. As Joel stated, it isn't the data rate that drops.

Since an 11b device can't hear OFDM (11g native) transmissions there has to be a way to make sure that the 11b devices don't interrupt the 11g transmissions. This is called "protection mode". While in protection, the transmitting device should precede any OFDM transmission with a CTS frame.

If you want to test this for yourself, you can use a program called NetPerSec to test actual interface speed. Shut down your 11b client and run an FTP download from a local machine (not the Internet). Then associate your 11b device to the AP and run the same test again. You will most likely see about a 40% drop in actual throughput.


Let us know if you need further explanation.

Hi Joel, Thanks for the response it was extremely informative, I hit a wall with what was happening, but you provided enough for me to figure it out. I 'm using a freeware program called QCheck by IXIA with provides a suite of throughput tools www.ixiacom.com. I'm logging client to client and now G client (54Mbps) to my WAP desktop which is on a 100Mbps link.
Sending 1000KB file the throughput is 15.779Mbps when the G client is operating at a WIndows XP tray indication of 54Mbps. WOW 8-O

If you have other throughput and testing tools that are demos or freeware please let me know I'd like to use them in my class for demonstration and learning. Also just a note that .5Mbps UDP stream works at .026Mbps only ouch!
Regards, Joe

[quote="joelb"]Wellllll... it's not the data rate that drops. It's the throughput. Your .11g clients will maintain the same data rate connection but, due to protection mode, their throughput will suffer when .11b clients associate to the AP the .11g clients are on. So, if you're just using the supplicant software, you'll never know you've got a .11b client in your .11g environment. You have to use a network throughput analysis tool to see that impact. There are many free tools out there that will show you this. Also, your supplicant software may have a "Site Survey" mode. If it does, you could run it and that would show the decrease in throughput.

Joel[/quote]

Nice Responses: I am learning more each visit.

Thought I'd mention this:

If you don't want to install freeware or shareware on every machine based on security concerns:

And if most of your machine users are Microsoft Windows clients; there is under the Control Panel and Administrative tools: Performance Monitor, better known as perfmon, that can give you basically the same thing.

This link talks about it:

http://www.wireless-center.net/Cisco-Wireless-Networking/Using-Perfmon.html

There is a handy Cisco Press book that goes through those steps above and has the graphics...if the link above, graphics don't load. I think you have to subscribe to view all the content, or buy the book, but it is worth the investment.

http://safari.ciscopress.com/158705227X/ch12lev1sec2#ch12lev1sec2