These sorts of technologies are really interesting, and will play a big part in getting the most out of limited wireless spectrum. But the big picture problem is that we're shoe-horning all this incredibly useful technology into the ghetto that is ISM band. We need more unlicensed spectrum, and we need better rules for that unlicensed spectrum that prohibit anti-social behavior: <a href="http://esd.mit.edu/WPS/esd-wp-2006-01.pdf" rel="nofollow">http://esd.mit.edu/WPS/esd-wp-2006-01.pdf</a>.
Link to technical paper: <a href="http://infoscience.epfl.ch/record/186014/files/saw-infocom13.pdf" rel="nofollow">http://infoscience.epfl.ch/record/186014/files/saw-infocom13...</a>
Hmm, I have not read his paper, but based on the article's mention of "divided into 13 channels" this is clearly speaking of the 2.4Ghz spectrum. Rather than try to distribute that evenly, you are FAR better off upgrading your AP and switching to a 5Ghz channel. First, there are far more potential channels in 5Ghz. Second, they are wider bandwidth (40Mhz, 80Mhz if 80211ac). And third, 5Ghz does not propagate through walls as well (a _feature_ when you have neighbors also using the spectrum).<p>My other thoughts: if he is using 8 channels instead of the standard 3 (1,6,11) then there will be some overlap; 80211 devices tends to better handle in-channel interference. <a href="http://en.wikipedia.org/wiki/List_of_WLAN_channels" rel="nofollow">http://en.wikipedia.org/wiki/List_of_WLAN_channels</a>
I'm glad someone at least seems to be addressing congestion. Every 802.11 standard seems to have hopped on the "multiple bandwidth by 2x" bandwagon, which is great if I live in a single family house with tons of space.<p>I don't, I live in a city, in an apartment building, surrounded by apartment buildings. I can frequently see 30+ networks. I'm sure there are wireless phones and other devices crowding the spectrum as well. Give me more channels, or better ways to share spectrum, anything other than dividing the number of usable channels with every update.
I actually had a similar idea recently but assumed it must have already been part of the protocol. Obviously it wasn't, and truly good ideas are never unique, so kudos to this guy for doing the work and coming up with a functional algorithm.<p>One question I have is how this would perform in an incredibly congested area like an apartment building. The author only modeled out 2-way conflict; I have no idea how this algorithm would react if placed in a situation with 15 competing Wi-Fi access points.<p>Overall a good, nerdy read - this is the kind of shit I keep coming to HN for.
Dynamic Frequency Selection and Transmit Power Control (part of 802.11h) is already a requirement for 5GHz in many countries: <a href="http://en.wikipedia.org/wiki/Channel_allocation_schemes#DCA_and_DFS" rel="nofollow">http://en.wikipedia.org/wiki/Channel_allocation_schemes#DCA_...</a>
When I chose a channel for my router, there was a clear option labelled "auto". If this algorithm for channel selection is novel, what does "auto" do?<p>Also, kudos for the idea. This is a much needed feature in today's crowded condos.