Hacker News Reader

Comment by callalex | Hacker News Reader

The WiFi alliance seems to be singularly focused on corporate/institutional deployments of WiFi while ignoring the needs of most individual consumers. I guess that makes sense because that’s where the money is.

The top priority of individual consumers is getting better range out of single-AP setups, which WiFi hasn’t really addressed at all since its inception. I guess MU-MIMO helps in theory with that, but doesn’t seem to in practice. With the exception of nerds on HN and really rich people with huge mansions, most people plug in one wireless AP placed wherever the cable/fiber/phone line comes into the house and want WiFi to have acceptable video conferencing 3 bedrooms away.

View on HN

I'm not sure there's really a solution to that issue though? There are strict limits on the channels you can utilize, and the power output of radios. There's no magic bullet to pass through physical objects.

I don’t deny that what I’m asking for is challenging, but if anyone would have the power and influence to pull it off it would be the WiFi alliance. Going higher and higher in the spectrum just sucks for the end user. Something in the 1-2 ghz range that isn’t a congested disaster like 2.4ghz would be extremely helpful for most home users. I understand that spectrum is hotly contested and rife with corruption in allocation processes, but it’s insane to me that the most important and most used wireless technology (by any metric really) is relegated to crappy leftover spectrum.

Why is that insane to you? The same technology that will allow your wireless signal to travel through all the walls in your house will allow your wireless to travel through all the walls in your neighbor's house too. The only reason wifi is even usable in a city is because of the limits in place today.

That’s not what I meant/said. It’s insane to me that we don’t just throw more spectrum at the problem. WiFi has definitively proven its utility and popularity. It’s insane that we allocate more spectrum to things hardly anyone uses instead of something that literally everyone is using, all the time.

The reason I think it's not that insane is because WiFi works pretty well as it is. I don't know anyone who seriously struggles with WiFi problems... those that did bought a (albeit expensive) mesh system.

The combo Wifi/modem that Xfinity gives out these days is also quite good, as much as I hate devices like that.

The truth is, WiFi is mainly a "problem" for nerds like me who LARP as network administrators and want faster-than-gigabit speeds in every corner of their home.

It’s holiday season, I still have a bunch of crappy quality calls with my family all the time. Talk to any doctor about their tele health experience, crappy WiFi gets in the way of them providing care coooonstantly. It’s WiFi that is the weakest link, the problems are rarely caused by the ISP.

There's so much nuance to the WiFi problem. So many put the ISP owned gateways on the floor, behind a TV, in the basement, in the attic, etc. Like an overwhelmingly large amount do not understand that putting your gateway in a bedroom closet on the opposite side of the house when you need it in say, the living room is nearly always the problem.

The only time I can recall when it wasn't for goofy reasons, was the launch of the Intel Puma chipset, where it would suffer from a CPU issue causing packet loss, or freeze until hard restarted.

More reading here: https://www.bleepingcomputer.com/news/hardware/lawsuit-brewi...

> It’s WiFi that is the weakest link...

Uh, I have a beef with this. I worked in a hospital (multiple medical facilities actually), and there are many, many medical locations have wifi poorly rolled out.

Hospitals are places that look nothing like a house from an electromagnetic perspective. If you don't put stuff in the right places, it won't work well. But if you do it right, then it works fine. And that's true of old buildings and new ones.

- - - -

EDIT - to clarify, it's not the wifi, it's the design of the wifi network (hotspot locations, backhaul, etc). It has nothing to do with wifi wireless technology itself.

That’s why I still use a desk phone at home.

We did just throw more spectrum at the problem. The 6GHz spectrum for Wifi 6E is a big change. It effectively eliminates the contention in higher bands. The downside is that need to buy new gear, and like all higher bands, need more AP.

More importantly, the same lower bands that let Wifi go through walls let mobile signals go through walls to enter your house. It is more valuable, both to the mobile providers and to the customers, to use that spectrum for mobile data.

Why do you think 1.2 GHz would be less congested?

Wi-Fi is congested because (1) lots of neighbors, especially in urban areas, and especially on lower bands where the signal penetrates walls better, and (2) unlicensed devices have to tolerate interference from other unlicensed devices, and (3) limited overall spectrum bandwidth, particularly as you go down the spectrum.

1.2 GHz would have all these same problems, but worse by being at a lower band.

There is consumer level software that lets you analyze bands and channels and select less noisy ones. If that kind of feature was somehow integrated into WiFi itself alongside widening spectrum bandwidth, that would probably go a long way. Easier said than done, I know, but there's got to be some better solution than running some program and looking at what's being used once a month or so.

I'd always assumed most access points already do this. For example, the 2.4GHz band has 14 channels and I've never told my setup which one to use.

>2.4GHz band has 14 channels and I've never told my setup which one to use.

This is an issue because some channels don't overlap, so if there's interference on a particular channel, your throughput is going to tank. Additionally, if the gateway/router/ap is always switching channels this will cause low power devices (nearly anything mobile) to disconnect and attempt to reconnect.

I'm still under the assumption that it's preferred to pull a Wifi survey before deploying any hardware so that it can be setup properly to function reliably.

Even my (extremely basic) modem from my cable provider does this. Auto channel switching/selection has been around for half a decade at the minimum.

My point is that there is so much other stuff sharing 2.4 ghz and it makes no sense to also have the most used wireless protocol also stuck in that same congested area. If we actually allocated spectrum in proportion to how people actually use it, WiFi should have a lot more available to use.

1-2 GHz is super congested. Some of the most popular mobile bands are in that range. With mobile carriers bidding billions of dollars for spectrum, there is no way to Wifi Alliance could find free spectrum. Also, spectrum is allocated per country; we are very lucky that the ISM bands are international.

Going higher in the spectrum means we can move as much as possible off off 2.4, and have less interference.

Mesh WiFi is common for consumers, and putting the mesh backhaul on higher frequencies frees up spectrum.

What we really need of course is like, another entire GB of low frequency ISM spectrum, though.

Seems like there's a bit of spectrum that could easily be freed up if they just took it from the commercial and public safety radios, which could probably use an upgrade anyway.

Spending a few billion giving them all spectrum efficient digital radios with a cellular fallback seems like it might be in the public's interest.

If they tried I would imagine they could provide make something so much more reliable, nobody would even think of complaining.

>Going higher in the spectrum means we can move as much as possible off off 2.4, and have less interference.

Correct me if I'm wrong but I recall the opposite. Lower frequencies pass through material better, and are less likely to be disrupted by ingress (foreign signal operating on the same frequency), where as higher frequencies have the ability to pass on more information. My previous time as a headend Coax/fiber technician has lead me to believe this. It's one of the reasons for example, Comcast utilized the lowest frequencies for it's upstream data pipeline, and the higher frequencies are for the downstream pipeline.

Agree on getting other systems off of the 2.4ghz band, but you still run into the throughput problem with that band.

Lower frequencies definitely pass through material better and generally have longer range, but with mesh systems you only need half or a third the range, because you've got repeaters.

5Ghz still goes through a few walls, kinda, mostly, so you can use it for closely spaced repeaters, kinda.

Which also means the 2.4ghz part can be turned down a lot, because it's only got to go to the nearest repeater, if you even need a 2.4ghz link at all for anything but IoT.

None of this applies to typical HN readers though, who all seem to be running an HFT stock exchange in their basement and need 10gbit at 0.5ms latency, but for the rest of us if it's good enough for 1080p Netflix and 720p zoom calls it's perfectly fine.

Except when the printer disappears despite everything having a strong signal because of mesh bugs... then it's not fine.

More wall penetration is not the solution. 900 MHz is the outcome of what you seek. It’s unlicensed so it’s filled with noise. And the noise carries far thanks to the low frequency. The entire band is only 28 MHz wide. “DECT 6.0” cordless phones use a 1.9 GHz band that is only 10 MHz wide.

There are very much ways to improve it. With no interference, there are a whole host of ways to improve range. Things like beam forming (sending the signal only in the direction of the client).

In busy WiFi environments like cities, there are probably even bigger gains to be had - tricks like CDMA and long FEC codes to allow your message to still be received whilst others are transmitting.

Beamforming and multiplexing (OFDMA) already exists in home wifi.

Beamforming like Ruckus stuff? Iirc I have ruckus AP’s that are well over a decade old that do it.

If they are able to increase the speed 4.8x, then instead of that they could divide that increased throughout into many more narrower channels of previous speed?

But you can't just increase strength of the AP. Isn't the issue the strength of the end device in sending a signal back to the AP?

Network comms are 2-way, it's not just a 1-way signal from the AP to your end device. It has to be able to talk back, and the signal a small battery operated device with a tiny internal antenna can send is only so good.

> Isn't the issue the strength of the end device in sending a signal back to the AP?

It absolutely is, you're correct. What most people fail to realize is nearly all mobile devices prioritize battery life first before anything else, including connectivity. On windows if you don't disable the power saving feature on your wireless NIC you're going to have drops, or straight up struggle to stay connected.

MacOS used to handle it better, but since BigSur, connectivity is about the same with Windows. I believe various Linux distros handle this better (like Kali lol)

If you truly believe this, please try to use 803.11a for daily needs.

I personally appreciate faster speeds, lower latency, better support for multiple devices, less contention, better support for wifi mesh etc.

Heck, I'm far more frustrated at wifi in 'corporate deployments' e.g. At a conference or hotel than I am with WiFi at home, so benefits there still improve my experience.

> I guess MU-MIMO helps in theory with that, but doesn’t seem to in practice.

I'd really like to see some thorough testing on this. My hunch is that mimo and more importantly beam-forming is something that works under lab conditions if every wifi device in range uses chips from the same manufacturer as the AP, and even then probably still depends on moon phase. Except maybe if you buy really expensive pro gear.

If some tech outlet would get a random assortment of popular APs and then test this stuff with a dozen phones and laptops that would surely be interesting.

Iirc tings like mimo and 1024, let alone 4k, qam don't work unless all clients can use it. So people who proudly say "I havent upgraded my phone is 5 years" are also the ones holding the wifi at 802.11n speeds

Low end and IoT devices also tend to only support older WiFi, so you may have to put them on another network.

MU-MIMO was not popularly supported (iPhones never did). It requires devices to be in different directions from the AP. https://www.smallnetbuilder.com/wireless/wireless-features/w...

MIMO is very useful. It literally multiplies link rate. https://www.wiisfi.com/#MIMO

Beamforming has marginal benefit and is real. Up close your device already has max speed. Far away beamforming won’t help much. It’s in the between that helps, and mostly in the downlink direction as beamforming requires the transmitter have more antennas and DSP resources (aka the access point that’s not running on battery and does not have to be miniaturized like a smartphone). But the uplink direction is usually the range limiting factor. There’s Maximal Ratio Combining for the uplink direction but it again is not universally supported in APs.

I'm not sure suburban and rural problems are as relevant as density in an apartment building. You can add APs, etc, but you can't stop collisions with neighbors.

> but you can't stop collisions with neighbors

I’m not convinced this is true. Right now all equipment screams at full power all the time, but it doesn’t have to be this way. They control the spec. They could enforce self-limiting when an AP detects more than 5 other APs, for example. And currently there is a lot of neat stuff in the spec for APs to collaborate with each other when they share a backhaul, but not so much for independent APs. Sure, you will have some jerks that go out of their way to override these features but for 99.999% of people who just plug in an appliance and never touch any configuration there are wins to be had there.

Just a shot in the dark hypothesis. Given the congestion is the issue and solely on the density of the population of AP's per given square meter (or cubic meter in apartments etc). I propose that AP's delegate the broadcast power based on real time bandwidth requirements based on upstream incoming packets to be retransmitted to the hosts either by the assumed internet interface or other attached to the AP. If the power is sufficient for transmitting the appropriate packets per second, given the bandwidth demands and latency requirements would it not increase available transmission spectrum for a collective dense environment?

WiFi 6 (ax) did add several new features designed to help with this, called spatial reuse.

Is there really a was without increasing power? The openly available frequency bands are limited quite a bit, so can the wifi alliance do anything here?

Yes, they can advocate for getting better spectrum allocated to WiFi. They would have a strong case since WiFi is by far and away the most used wireless technology it’s just that nobody is really advocating for it to politicians in any meaningful way.

I got over that using WiFi enabled power line adapters.

Reply on news.ycombinator.com