drinkingbird
Part of the Furniture
OK, fine, you made me look it up from the testing years ago. The issue was Aruba and Cisco AC APs, along with some Intel NICs, were sending "40mhz intolerant" flag on 5Ghz channels (something that was only designed for 2.4Ghz but for some reason could be enabled on 5Ghz with those vendors, and probably others that we did not test). Clients should theoretically still have been able to connect at 20, but 40mhz clients would not connect at all when that was set. 80mhz clients connected fine. It was the default setting on early versions of code on at least one of the Aruba APs when you hardcoded the width.
The original discussion was around a user having issues with some devices connecting. It was suggested by someone to hardcode the channel to make it "MORE" compatible. I suggested that it was actually the opposite, letting it stay at auto was more compatible and I had seen issues with clients connecting when it was hardcoded. Yes I probably misspoke by saying it WOULD reject clients that didn't support that width, was poor recollection of testing that was done years ago, when the correct statement would have been SOME APs MAY reject clients not matching that channel width. Chalk it up to WIFI not being anything to do with my role for years now, other than what I do at home. But, considering the user's issue, I stand by my recommendation to not hardcode the width.
As far as whatever google examples of 80mhz AP with 40mhz client, that was never the discussion, the discussion was around hardcoding it to xxmhz vs leaving it in auto/compatibility (or as Cisco calls it, "Best"). Never disagreed that an AP running at 80 could handle 20 or 40 mhz clients, that has always been by design and never seen one with auto width have any issues with various client widths at the same time.
Showing 1 example from 1 model of 1 brand home Asus router is not "Proof", as is clearly the case here. But you're the one who claims you have infinite enterprise experience across all devices and not so much with the Asus, yet you use the Asus as your concrete proof.
And no, Cisco does not always follow the standards, that's been painfully obvious for decades. CGMP, ISL Trunking, Phantom power, could go on for pages and pages. Their M.O. is release their own standard first then support the ratified standard at a later time (and even then, often with their own interpretation of it). They still insist, even in the latest versions of code, on enabling their interpretation of 802.1Z draft on all gig ports by default when it was never even ratified and eventually cancelled.
My stance on employee retention is different from yours too, the ones with attitude problems get let go. At least someone who doesn't know something can be taught.
I'm sure you'll need to get the last word but I'm not getting dragged any further down this hole.
Last edited:
