Lee, We're running wide channels only in 5 GHz. We experimented with wide channels before going live and found no adverse side-effects with our 11a users, and all of our 11n capable clients we're happy too.
The big driver for wide-channels was that while we have Gigabit to the pillow, most folk have abandoned it for WiFi. For those users that are in our digital media programs, the higher max throughput is a necessity as they tend to move a large amount of data back and forth. As for verified throughput, the broadcom-based chipsets seem to be a bit iffy when it comes to repeatable performance, but the atheros-based chipsets seem to scream along. I can attain 130-140 Mb/s consistently. If I recall, you're a Cisco shop. There is a performance "issue" that we originally experienced which did not crop up unless wide channels were enabled. This had to do with the way packets are fragmented within the LWAPP protocol, and could result in some less than stellar data rates for 5GHz 11n with wide channels enabled. I believe this is fixed in the 5.2.x code-base as it now uses CAPWAP. If you're not on 5.2, the solution was to add the following to the routed interface that your wireless users were dropped in to "ip tcp adjust-mss 1400" There are also a few issues with Mac's that have broadcom-based chips. They tend to have very inconsistent data rates when using 5 GHz channels in the UNI-IIe bands. I solved the issue by turning those channels off (for now), and just using the following: 36, 40, 44, 48, 149, 153, 157, 161. As for the higher data rates versus any sort of per-user throttling... For the past five years, our student's wired connections have been Gigabit, yet I've never had someone come to me to complain because they weren't getting that rate in/out to the Internet. I think folk understand that what their computer reports is not what they will actually get. We've been live on our new Cisco system since the beginning of our Academic year, all Cisco 1252 11n radios. Nothing but praise and thanks from our community. best, Jeffrey D Sessler Director Information Technology Scripps College >>> Lee H Badman <[EMAIL PROTECTED]> 12/1/2008 6:48 AM >>> Knowing that some have gone before us on taking the 11n plunge... we are very much just starting to play and plan. Some questions, if anyone feels like sharing there experiences. - It seems like universally, wide channels in 2.4 GHz is deemed a bad idea. But for those who choose to wide channels in the 5 GHz side of 11n- o What is your driver for using wide channels? Are there any goals beyond seeing higher data rates (specific applications, etc.)? o Where higher data rates are seen (like the client says 300 Mbps)- are you seeing fairly consistent verified throughput? (mine seems to be all over the place in early test with multiple client devices) o Has anyone been challenged on disenfranchising 11a users that can't due wide channels when compatibility is not allowed, and how do you respond? - For Ekahau and AirMagnet Surveyor customers that are using these for 11n planning (I have and know both products)- are you finding the accuracy of modeling to be acceptable with the 11n variants? - Does anyone struggle at all with the new higher data rates versus any sort of per-user throttling at your campus Internet edge, and how do you communicate why the higher data rates can be misleading for Internet traffic that has known bottlenecks? - In general, have you found any glaring problems with going to 11n, in any operational mode? As always, on list or off is fine, and this is not an invite for vendors to call me. Regards- Lee Lee H. Badman Wireless/Network Engineer Information Technology and Services Syracuse University 315 443-3003 ********** Participation and subscription information for this EDUCAUSE Constituent Group discussion list can be found at http://www.educause.edu/groups/. ********** Participation and subscription information for this EDUCAUSE Constituent Group discussion list can be found at http://www.educause.edu/groups/.
