Ah, something I know well enough to comment on.

802.11b has a theoretical max /throughput/ of about 5-6 mbps (this is somewhat arguable, but I have never seen a real world 802.11b device break 5.5 mbps). You are right that 200 KBps works out to about 1.6 mbps and if it is reporting a link speed of 11 mbps then there is definately something wrong.

I would make sure your testing is accurate though. The Aironet (IIRC) should have a bit speed report you can run expressing the five minute average load on an interface. Something like "sh int <interface>" shoiuld give you a actual bit rate. Make sure no one else is using the link while you are testing as well. Watch for periodic interference on the link (cordless phones, video cameras, microwave overns, etc) by watching the signal and noise graphs (not just the SNR). Graph them over the time of the transfer if possible (I don't remember much about the Cisco site survey tools, after we eval'ed them we went with someone else). Periodic interference is the hardest thing to track down. Make sure this is a point to point arangement and not a point to multipoint or something else.

Make sure there are not overlapping channels, particularly since the only non-overlapping channels are 1, 6, and 11 (and even these can overlap if the equipment isn't perfectly tuned [Cisco is pretty good that way though]). Watch for wireless clients trying to hop onto your point to point link also.

Just some additional information for the list, 802.11a has a max theoretical throughput of about 27-30 mbps on a max link speed of 54 mbps. 802.11g (untested, this is just from chipset reports I have seen), peaks out at about 20-25 mbps (IIRC, this isn't NDA now that products are announced/released).

Wireless is a lot more tricky than wired, since you have to account for a lot more evironmental variables and there are no "full wire speed" devices out there :-)

Adam

PS- If anyone knows Friganoid the Wardriver I am looking for him/her/it/them.



John Nielsen wrote:

This isn't directly unix-related, but I know a lot of you have more experience with wireless networking technologies than I do.

A friend of mine asked me to help troubleshoot his network. He is using several Cisco Aironet 350 bridges (100baseT/802.11b) to connect different network segments. There is frequently a lot of latency on the network, and the wireless links seem to be the bottleneck. The bridges report that they are operating at 11Mbit/s, and have 85% signal strength (and 95% signal quality). Yet transfer rates seem to be topping out at about 200KB/sec. Even allowing for a lot of protocol overhead, etc., that can't be more than about 2 Mbit/s.

I'll probably get on the phone with Cisco tomorrow, but I wanted to get some real-world experience from the list. What kind of transfer rates can I realistically expect over an 802.11b link? Is the correlation between signal strength and throughput roughly linear or how does that work? Any general suggestions on how I might improve matters?

Thanks,

-John

____________________
BYU Unix Users Group http://uug.byu.edu/ ___________________________________________________________________
List Info: http://phantom.byu.edu/cgi-bin/mailman/listinfo/uug-list



____________________
BYU Unix Users Group http://uug.byu.edu/ ___________________________________________________________________
List Info: http://phantom.byu.edu/cgi-bin/mailman/listinfo/uug-list

Reply via email to