I live in London (my neighbours are a LOT closer than 1/4 mile) with a shitload of devices on my WiFi and I can still happily stream 4K TV while my my entire family uses iPads or laptops. Am I breaking the laws of physics?

>Am I breaking the laws of physics?

No, you're just using subjective feelings rather than objective measurements. 4K doesn't specify a particular bit rate. If there is going to be a productive discussion, it needs to be acknowledged that there is a huge difference between streaming 1080p and Bluray 1080p.

Why would anyone stream a Bluray?

No one would want to stream a Bluray. The point was to highlight that there are very different things that are still described by "1080p" and what is more important is the bit rate and codec rather than the common 4K, 1080p, 720p, or even "HD video" names.

I don't know why you're even talking about 1080p, though I think I vaguely get what you're trying to say.

To play devil'a advocate, a Ruckus Zoneflex R710 ought to work well in such places.

Also, I have gotten wifi working well in non-rural areas using Ubiquiti Unifi AC Lite APs. In suburban New York, a couple of them blanket my home. In my grandmother's Shanghai apartment in a building that is a mix of concrete and plaster, just 1 is sufficient (although for a much smaller volume), despite dozens of nearby APs. Admittedly, I moved to 5GHz there because I was literally the only one using it, but 2.4GHz had worked fine when I tried it. Peak throughout is higher on 5GHz due to higher channel widths though.

I might make an analogy: "I use a dog whistle and no matter how loud people are talking, my dog always hears it"

OK, now take your dog to a place full of dogs and other dog whistles. See how well it works.

Once the spectrum gets noisy, interference happens. Just because you found a quiet spot in the spectrum doesn't mean that spectrum sharing isn't a physical reality anymore. It just means you found some to keep all to yourself! Until someone else starts using 5GHz too.

I said that I was able to use 2.4GHz reliably despite it being crowded. I only switched to 5GHz because I noticed no one was using it and I saw higher bandwidth on it from 5GHz 802.11ac supporting higher channel widths than 2.4GHz 802.11n supports.

In fairness, I had made tweaks to iwlwifi's kernel module options in order to have usable wifi on my laptop. Until a year ago, I was accustomed to have a hellish experience in congested areas until I turned off Bluetooth coexistence support in the iwlwifi driver. That feature would almost always cause severe packet loss on my T520 in crowded environments even when there did not appear to be any Bluetooth traffic.

Anyway, it is possible to get wifi working well for certain workloads in crowded environments. Making it work well in general might require better drivers and better equipment than one might have at first though. For example, I understand that ath9k was a disaster when it was first made, but it is fairly decent now. I ran an AP off an ath9k USB wifi dongle in China at one point and it worked well too. I doubt that would have worked as well with drivers from 5 years ago.

Also, from what I understand, a used Ruckus Zoneflex 7982 off eBay ought to be able to handle just about any environment fairly well. Their proprietary beam forming hardware is special because it attenuates signals coming from other directions than that of the client. The only exception from what I have read is when other wifi equipment is right next to it (e.g. practically touching). Some review of it said that it failed to work until they moved other equipment away after consulting Ruckus. I cannot find a link to it though.

I placed an order with an eBay merchant that I expect to receive soon. $90 per used Zoneflex 7982 (that had a $1099 MSRP when new) is a bargain considering that these still provide some of the best wifi in the world. There are reports of these getting good throughout through multiple concrete walls:

http://www.cio.com/article/2388827/wireless-networking/revie...

A newer version of the midrange model in their product line (that has inferior radio specifications to the top end model) was able to give a cell phone decent throughput from 225 yards away when the (omnidirectional) AP was in doors:

http://www.evdoinfo.com/content/view/5077/64/

As far as I know, the only better access points are the newer Ruckus models that replaced it. I am really excited to be getting one later this week. :)

For a large fraction of the population, this post might as well be an advert for Apple-style plug-and-play routers.

Why? I was talking about APs, not routers.

If you meant to imply that I am somehow paid to say such things, I assure that I am not. I am a fairly well known OSS developer and such a thing would be damaging to my reputation. I just happen to be genuinely excited about this topic for the first time in years at this moment.

I just returned from visiting my grandmother in China, who is in the hospital. I stayed at the concrete and plaster apartment complex where my grandmother's apartment is. The horrible pentetration of wifi signals made me think plenty about wifi and that eventually lead me to discover that Ruckus's old equipment would work far better and is selling used on eBay at pricepoints seen in consumer grade gear. There really is not much benefit to Ruckus from the free advertisement of me talking about the used Ruckus equipment on eBay. They are not making any money off the used hardware market while the pricing for new Ruckus equipment is so high that it might as well not exist as far as the majority of people are concerned.

I was not clear enough.

You managed to achieve good network performance by using your specialized skills.

Most people don't have the skills and aren't interested in learning them. This is the market for Apple's APs. The interesting description of what you did to get stuff working is close to meaningless to most people, and hence by contrast makes the trusted-brand easy-peasy gear look even more valuable.

I've wrestled with wireless gear since before 802.11 existed, including writing my own radio firmware, and I have an Airport at home which in practice means I don't need to do any of that awful stuff any more. In my lab we throw away about half the APs we buy due to mystery incompatibilities. Radio is still hard compared to most other things.

Point taken, although a minor counterpoint is that I do all that I can to share my knowledge whenever people ask. Last Sunday, I had advised a complete stranger on how to get cheaper cell phone service. We had started talking after I asked if it was alright if I joined his table at a crowded food court. Upon finding out that I am a software engineer, he started telling me about his problems and I did what I could to help him. It was a nice use of part of my Sunday. :)

Random thought: is there a reason we couldn't construct apartment buildings such that each tennant had a Faraday mesh embedded in their shared walls, but not in exterior-facing walls?

It would do strange things to cellular and broadcast radio, but it wouldn't destroy them, I think.

IMO, cost is a big factor here. Adding things like insulation, EM sheilding, or anything else other than the minimum needed water water and electrical lines to the inside of walls raises the cost of construction. So if it isn't required by the function of the room or building code, you're unlikely to see such fanciness, unless it's some high-end luxury place where the shielding is a selling point. Expect higher rent in such a place to cover the additional costs of constructing those fancy EM-shielded units.

IANAP but I would think such a cage would cause WiFi to reflect internally, causing as much if not more trouble than noisy neighbors.

The "real" solution is for people in dense apartment buildings to just turn their damn Tx power down, but that's a prisoner's dilemma if I've ever seen one.

I actually tried to approach this problem in my own apartment complex. The effort died almost immediately. Very few consumer wireless access points even expose Tx power as a configuration parameter at all, even buried in an "Advanced Settings" section somewhere. You're lucky if they even expose channel selection at all.

So with the craptastic firemware of most consumer wifi APs, the AP hops on channels 1, 6 or 11, and stays there, screaming at full power. And the consumer, even if they know enough about radio to realize such is suboptimal, can't do anything about it without changing the firmware or buying a more expensive device.

From my experiences, lots of these consumer APs frequently end up landing somewhere like channel 4 or channel 9. :/

In a really busy area, what's the issue there? If there are tons of people on 1/6/11, you may as well hop into the middle bits - there's already plenty of interference going on in the usual channels.

With 802.11g you're actually taking up four channels out of the available 12? So people hopping in the middle are messing with people on both sides of them. Sorry for the mobile link. https://en.m.wikipedia.org/wiki/IEEE_802.11g-2003

However, if the 'standard' channels are already congested, people are already messing with each other. Look at the semicircle graph on your link - someone on channel 4 is messing a fair bit with someone on channel 5, and a little bit on 6. However, someone on channel 6 is messing a lot with another person on 6.

So, people on the same channel silence each other (CTS/RTS), but on other channels it is simply seen as noise. If there are lots of people on 1/6/11, you're going to spend a lot of time silenced.

It's one of those things that you just have to try out in your own particular circumstances. Will you get better performance if frequently silenced, or better performance dealing with a high level of 'noise'?

Actually WiFi was designed from the very beginning to do well despite reflections. My thesis advisor knew somebody who was applying ideas from "quantum chaos" to make WiFi robust to various environments at Bell Labs, back in the 1990s, before Carly Fiorina bankrupted Lucent.

It's a prisoner's dilemma in the case where everyone controls their own wireless APs, sure.

In the case where the building is served by one major ISP, and they provide combination gateway+APs, I could see the ISP deciding to "solve" the problem by turning the whole building into one AP mesh (with only one AP actually active per unit sphere) where each client gets a VLAN with tagged QoSing over the last-mile "backplane" back to their own gateway.

Though, if you're going that far, you may as well just call yourself a WiMAX provider and tell people their gateways are just extension picocells for the network, rather than end-user equipment. Which is basically what you get already from the ISPs that offer "public wi-fi for customers at our hotspots" (i.e. broadcast by all APs on our business plan), except that said wi-fi would be the only product, and would be served by residential gateways as well.

There is wifi blocking paint and wallpaper. http://www.lessemf.com/paint.html for example. That said, this may not be desirable if you have a cellphone without a local nanocell or voip switching.

There is wifi blocking paint and wallpaper. http://www.lessemf.com/paint.html for example. That said, this may not be desirable if you have a cellphone without a local nanocell or voip switching.