ac is a new protocol that not only is capable of gaining more speeds by the use of the extra channel bandwidth, it also has interference avoiding technology build in and by making use of the 256QAM rate that even on a 20Mhz channel should give you more throughput.
So, even in congested spectrum situations AC will benefit. Maybe not up to the highest level the technology theoretically promises, but still better than a 'n' radio in the same situation.
What we really need now is a manufacturer that comes with a good, all in one, 3 chain high gain (+28dBi) antenna. Than we are in wifi walhalla!
(Well, at least for a while.....)
What interference avoidance technology is built into the actual Atheros Wave2 Chipsets?
The standard does not claim higher near/far channel rejection than 11n. The vendors do
not provide this numbers ...
I must say I haven't found the article yet that I red some time ago. But what I can remember was that while ac is using 80 or 160 wide channels, within these it will try to communicate with its other radio on only these channels that are free of signals from other 'legacy' 20/40Mhz channel using radios.
It sort of 'bypasses' these channels to use only the ones with the low noise. And this on a dynamic base.
What I can remember this was one of the mayor advantages of ac if it comes to working in frequency band overlapping circumstances where ac was thus able to choose the best channels available in its spread. Of course this would mean lesser throughput because some spatial stream were rendered 'not usable' but since it happened on a dynamic basis these omitted channels would be used again when the alien radio signals would have gone...
Actually it came with a scheme to show that in a 160Mhz spread where two legacy a (or 40Mhz 'n') channels within this same spread were used by other radios these channels were just not used by 'ac' radio. This would mean a degrease in total maximum throughput for the 'ac' setup but the still usable channels were still able to transport considerably more traffic than a 'n' (or a 'a') radio would be able to. Because it could still use two, three or more usable channels.
Basically in an area where the whole 5Ghz radio spectrum would be used by legacy 'a' or 'n' channels you could still put several 80 or 160Mhz wide channels in the same region and the ac protocol would still be able to deliver more speed than a 'n' would be able to do.
For building backhauls this would be ideal. You could make a 160Mhz bandwidth wide link and if some of the underlying frequencies would see some interferences (at times) these were just not used but the rest of the traffic would still be steered over the other available 'quiet' channels.
We all know that interferences at times are very dynamic too. This way if using a 160Mhz wide channel where at times it 'sees' interferences on two 40Mhz channels we would still have 2 other 40Mhz channels left to transport our traffic. That is still 200% of what a 40Mhz 'n' channel could deliver!
And to be honest, if this would not be the case the whole 'ac' history with its high throughputs and 'the need to have' factor would be an empty air bubble whatsoever. If it wasn't able to use the extra wide channels due overlapping/interferences with other legacy 20 or 40Mhz radios 'ac' would be almost completely useless in the 'real' world..... Spectrum is used and stretched to its extend almost everywhere. It would make no sense to design a new protocol that would not come with one or another system to handle the congested spectrum of nowadays.....
Tonight when I can find some time I will try again to find that specific document.