On 3/17/2014 10:56 PM, Frank Sweetser wrote: > Well damn. I had been hoping that clients would have gotten at least > a little smarter in their roaming decisions, but clearly that was just > wishful thinking.
Well, some client drivers will let you tweak the "advanced" settings. I know I can "prefer 5Ghz" and I can "prefer N" on my aging laptop (yeah, it was my Dell below, trying to track a client issue back to the wiring closet where we hung a leftover Aruba AP65 (a/b/g only), and my own freaking laptop was hanging on to an "n" on a floor below). It also appears that iOS 7.1 update changed the "captive portal detection" on Apple devices... so if you're having issues with your "registration portal" for new devices, you might double-check their captive portal site check. We've had a *flood* of iOS devices that "couldn't register on the network" today... Jeff > > Thanks all for the confirmation... > > Frank Sweetser fs at wpi.edu | For every problem, there is a > solution that > Manager of Network Operations | is simple, elegant, and wrong. > Worcester Polytechnic Institute | - HL Mencken > > On 3/16/2014 1:04 AM, Jeff Kell wrote: >> Have seen similar results with Dell laptop locking onto 802.11n at a >> distance >> and ignoring "same room" a/b/g. We are trying to avoid mixed >> deployments, and >> sounds like the same concerns extend to 11ac as well. >> >> Jeff >> >> On 3/15/2014 11:12 PM, Alok Vimawala wrote: >>> Hi Frank, >>> >>> We just had an interesting incident in one of our buildings where >>> half of >>> the ac radios stopped working. The building has Cisco 3602i APs with >>> the >>> add-on 802.11ac Wave-1 module. So, the building turned into a mixed >>> 802.11n >>> and 802.11ac deployment on the 5GHz spectrum. What we saw in that >>> building >>> was that new Apple MacBook Pros with the 802.11ac capable chipsets were >>> preferring to associated with a bad 802.11ac signal rather than >>> connecting >>> to a great (AP right above the laptop) 802.11n signal. >>> >>> Clients seem to prefer protocols with highest theoretical throughput >>> regardless of signal strength and that behavior hasn't really >>> changed since >>> the days when 802.11n was first introduced. My recommendation would >>> be to >>> avoid mixed 5GHz 802.11n and 802.11ac environments. >>> >>> Thanks, >>> >>> Alok Vimawala >>> University of Michigan >>> >>> >>> On Sat, Mar 15, 2014 at 9:54 PM, Frank Sweetser <[email protected] >>> <mailto:[email protected]>> wrote: >>> >>> Hello all, >>> >>> we're beginning plans to upgrade our wireless infrastructure >>> from 11n >>> to 11ac, and I'm hoping that someone can chime in on their >>> experience >>> with mixed capability buildings. >>> >>> When we first went from 11a/b/g to 11n, we found that clients in >>> buildings with mixed capability APs had some odd roaming issues >>> - and by >>> "odd", I mean utterly braindead. A fair number of clients would >>> aggressively latch onto an 11n AP at -80, while ignoring an >>> a/b/g AP in >>> the same room at -50, with predictably poor results. In the >>> end, we had >>> to ensure that buildings were upgraded in full, rather than >>> incrementally, to fix the complaints. >>> >>> My question is, has anyone seen similar issues in buildings with >>> a mix >>> of 11ac and 11n APs? >>> >>> -- >>> Frank Sweetser fs at wpi.edu <http://wpi.edu> | For every >>> problem, >>> there is a solution that >>> Manager of Network Operations | is simple, elegant, and wrong. >>> Worcester Polytechnic Institute | - HL Mencken >>> >>> ********** >>> 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/. >>> >> >> ********** 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/. > ********** Participation and subscription information for this EDUCAUSE Constituent Group discussion list can be found at http://www.educause.edu/groups/.
