Hacker News new | past | comments | ask | show | jobs | submit login
Introducing a new kind of Wi-Fi system (blog.google)
158 points by rmanalan on Oct 4, 2016 | hide | past | favorite | 192 comments

Just for everyone focusing on gimmicky features, this is an OnHub, which are intended to just work.

Under the hood, they run ChromiumOS, including things like trusted boot (yes, this home router has a tpm), and the A/B partitions so when you get an automatic update, (which happens during a time of low bandwidth usage), it reboots into the new version in about 6 seconds. The security team is awesome: they pushed an update to all onhubs within 48 hours of public disclosure of a critical RCE earlier this year.

There has been some slowness to expected features, like ipv6, but the PMs have been clear about their goal: they won't include a feature that's buggy. It's exactly what you want in a piece of infrastructure.

Wait, OnHub doesn't support IPv6?! That's inexcusable. There's no reason IPv6 couldn't "just work" too. Google Wifi had better support IPv6 at launch.

If you require IPv6 so crucially, you probably shouldn't be using consumer grade gear.

I've been using IPv6 on a consumer-grade cable modem connection with a Cisco DPC3008 modem and a previous-generation Apple Airport extreme since 2012. There weren't even any difficulties getting it working.

There aren't any difficulties if you don't get it working either (yet). The Internet still works with no issues.

There are quite a few providers in the world that run Dual-Stack Lite (native IPv6, IPv4 is tunneled to provider-run NAT gateways) and they all hand out consumer-grade gear.

Since the OnHub only comes with Ethernet you'd have to put something between it and the wire anyways of course, but it then has to be more than a dumb modem, and you'd be forcing all traffic through the tunnel.

IPv6 isn't without problems, but many people use it without noticing, through consumer devices way cheaper than this. The ones complaining the most are the enthusiasts which suddenly can't reach their home server via IPv4 anymore.

My ~$100 Linksys router supports IPv6. What's Google's excuse?

Yet another device Google will sell, and then discontinue in 24-36 months.

> Yet another device Google will sell, and then discontinue in 24-36 months.

You don't know that.

Besides acquired products (Nest) and Android phones, Google has a pretty poor track record of long term hardware support. I can still get firmware updates for my 4 year old Linksys router.

I've been burned, and I refuse to purchase yet another Google device for a market they're just testing the waters in. I'd rather spend my money with someone who needs it to support their product, and that product is their business.

To be clear, you're saying they're just testing the waters with Chromium? Or with Fiber, which makes its own hardware STB/routers, or with this whole internet performance thing in general?

Chrome devices have 5 years of updates guaranteed. I have a cr-48 from 2010 which still gets updates (currently on -dev 56). How many times has there been a story on HN about RCEs unpatched for months or years in consumer routing gear? The fact that you purchased that hardware doesn't seem to be sufficient incentive based on the record. But when you have an incentive of a secure and performant internet, and also have the talent, infrastructure, and existing codebase to make something work and work well, you have economies of scope that make supporting it much easier and cheaper.

IPv6 is for everyone: consumers, businesses, whatever else exchanges packets.

That kind of thinking is 10 years out of date. Consumer broadband routers sent to subscriber homes for free support IPv6.

Gotta love easily NSA-dioded consumer gear which falls over under load, reboots randomly and doesn't play nice with standards.

Really, it doesn't do IPv6?

How could they launch such a faulty product. If it was 3 years ago maybe, but it's not really optional now.

>That's inexcusable

Except it's not, we live in an IPv4 world.

Edit: How are you guys downvoting me on your v6 connections? HN only supports v4.

Maybe you haven't checked lately but we live in a world where IPv6 adoption is over 10% (over 25% in the US) after doubling each year for the past 6 years. https://www.google.com/intl/en/ipv6/statistics.html

It's on the roadmap, but so far hasn't passed QA testing.

The engineering time is going into upstream too, so ath9k, hostap, and other big networking libs are benefiting from this.


Jan 2016 -> 10% Jan 2017 -> 20% Jan 2018 -> 40% Jan 2019 -> 80% Jan 2020 -> 100%

Seems quite good to me.

Why does IPv6 availability matter here? The amount of users that need, want or even directly benefit from IPv6 is vanishingly small.

I disagree. Directly, yes, the average user doesn't care. But application developers care; peer-to-peer protocols become a lot trickier with IPv4 due to the pervasive deployment of NAT; two machines ostensibly on the Internet can't connect to each other, requiring instead the use of STUN servers, which then requires infrastructure somewhere, or just doing it client-server, or some mix like having "supernodes" (like Skype, prior to MS tearing it out) that route traffic for NAT'd devices.

The ability to actually connect arbitrary devices, I hope, will be something that people will take advantage of. I know for many game servers I set up with siblings, the ability to not need to mess with a router's crappy "port forwarding" would be a welcome change. (Even if I had to mess w/ some local firewall, but that can perhaps be much more tightly integrated or at least, a better UX.)

Yes, but until you have the percentage of overall IPv4 usage down to less than 5%, 10%, or even being generous say 15%, developers will still have to deal with those things (NATs, STUN, TURN, etc) anyway.

IPv6 has been around for almost 20 years now, and is only recently cracking 10% (and I wonder how much of that 10+% is also dual stack). IPv4 sure as hell isn't going away in my lifetime. Who knows, maybe the lifetime of my kids too. What a mess!

Because the quicker we adopt it the better.

The shittiest of routers support it so when you get one from a major internet company you should expect that it has support for an internet protocol which has been out for 18 years.

Doesn't qualify as "inexcusable" for me, sorry.

IPv6 is important, but it's not like your ISP will lower your bill if you forego IPv4, and it's not like there are IPv6-only sites. If I recall correctly, I have to pay _extra_ for an IPv6 address.

IPv4 addresses have been more expensive than IPv6 subnets on every dedicated uplink I've gotten prices for over the last two years. In fact, an IPv6 /56 is usually free, or cheap enough to be effectively free since the fee is mainly a NRC for the time to set up the route, if you ask for it and the provider supports IPv6. In contrast, IPv4 addresses often incur a MRC based on the number of usable addresses you request.

> If I recall correctly, I have to pay _extra_ for an IPv6 address.

That's weird, here we get a /48 v6 block by default and one /32 v4 address. You'll never need more v6 addresses, but each v4 address comes at a monthly fee.

Most of my connections are over ipv6.

If this doesn't support ipv6 then it can only connect to legacy sites. Yes that's pretty much all sites today but this is inexcusable in a new product.

Well you can never upgrade to a IPv6 world if you don't add it to routers...

Ah yes, deprioritize features then spin it as "we just want to make sure it's perfect", as if IPv6 is some kind of work of art or delicate baking.

Just want to correct a little thing: TPM is used for measured boot, trusted boot is something else

Ah, you're right. I was using the term loosely.

What are the user privacy policies and technical implementations?

Lots of vendors already offer this. It's pretty standard for commercial WiFi units. It's just slightly more expensive than dumb routers.

* Eero: https://eero.com/

* Ubuquiti: https://www.ubnt.com

* Cisco: http://www.cisco.com (sized for larger buildings)

Does this new Google device come with the Google Fi feature which backhauls all your traffic to Google via a VPN?

For Ubiquiti I'd suggest linking to https://www.amplifi.com instead of ubnt.com. AmpliFi is their mesh networking product aimed at consumers, whereas almost everything on ubnt.com is aimed at enterprise use.

With Amplifi out, until I see performance numbers, I'm still going to suggest UniFi AP AC Pros due to how fucking ridiculously good they are.

I can't speak to Amplifi as I've just heard about it a few moments ago, but I second the UniFi AC Pros. Long story short, the 2.4ghz spectrum where I live is saturated, and the 5ghz spectrum is wide open. 2 APs, a Cloud Key, and a PoE switch later and I have full 5ghz coverage on both floors. I can bury the capacity with file transfers, BitTorrent, Netflix, etc and it doesn't skip a beat.

The bad news is if you're used to the cost of consumer gear, this ain't it. When it was all said and done it set me back $400. Throw in the cost of a pfSense router (because my old workhorse died recently) and we're at about $600. So, much like everything else, getting to that first 90% is cheap. That last 10%, not so much.

I have a UAP-AC-Pro and a second UAP-AC as an extender, and they're great. But I ordered an AmpliFi HD anyway because it will probably be easier to maintain and I'm interested in the mesh networking. I'm hoping it lives up to expectations.

No. It comes with pretty much nothing but a nice little price break. You get a 3-pack for $299. The Eero 3-pack is $499.00. This is a serious price break in this market, which is currently over-priced. You can get 3 Ubiquities for almost $300. A lot less user friendly, but $500 for 3 units is holding back mass adoption.

Personally, for my home I need just two units. Shame there's no $199 2-pack yet. It would be an instant buy for me.

Ubiquiti's Amplifi is $199 for 3 units.

Amplifi has 3 models. The high end, the amplifi HD is $350 for a 3 pack

If they made a 2 pack it would probably be around $230 with my guesstimate.

  3-pack $300 or $100/item
  2-pack $230 or $115/item
  1-pack $130 or $130/item

Buy the three pack and sell one?

Also Portal Wifi: https://portalwifi.com/

This was started as a kickstarter campaign and they have just begun shipping out the first units.

Like the Google APs, Portal can form a mesh and automatically figure out which channels have the least interference.

I have quite a few of their routers. Very underrated.

Yes. That cloud setup punches through firewalls and is just turn key. I wish they made routers.

Little known, you can do wireless extending (main + one hop end nodes + clients) or WDS mesh (main + one hop relay + one hop end nodes + clients) with Airport:

Overview : https://support.apple.com/en-us/HT202076

Extending: https://support.apple.com/en-us/HT204617

WDS: http://support.apple.com/kb/HT4262

While the wireless distribution mode drops speed (as does Ubiquiti etc.) it's still faster than a DoCIS 3 cable modem, and the Airport Express for end nodes is cheap.

> Does this new Google device come with the Google Fi feature which backhauls all your traffic to Google via a VPN?

Not applicable.

The purpose of Google Connectivity Services is to provide protected access (encrypted via VPN) to the Internet on an unsecured 802.11 network like one that you'd encounter at a coffee shop.

In that case, GCS sets up a tunnel from your phone to Google's edge network, traffic from your phone egresses via an IP address dedicated to GCS and then you're surfing the Interwebs.

If you've entered a SSID and a related PSK or WEP password (as one typically does on a home network), GCS isn't available to use.

Use of GCS is optional. It can be disabled as desired.

Big shout-out for Eero (https://eero.com/), which is the best WiFi product I've ever used.

I'm skeptical about the flux capacitation feature they list. The rest looks neat.

Serious question: whose WiFi is not working, to the degree that they think, "I really need to get a modern router to make this Internet thing actually work."

I'm seeing about a dozen fancy modern wifi routers all trying to solve a problem I'm not sure exists.

I had nothing but problems with typical consumer hardware. Having to restart routers all the time. Dead spots. When family would visit (both my wife and I have large families) they would connect at least 2 devices each to my wifi (tablet, phone, laptop, etc). My wireless would just tank. They would always ask me why my internet was so terrible.

Last year I bought an Edge router and a higher end Netgear hotspot. Works wonders separating things out and letting me easily add separate "Guest" wireless networks on different channels or whatever using my older devices.

If I were to have done that change now I would've definitely used a device like Google's. But to do something like that would've been outside my budget. I wish I could easily add new hotspots that just seamlessly worked so that I could eliminate the dead spots on the edges of my house and in my yard over time.

Maybe there's something I don't know already and I don't need Google's new Wifi System. But it definitely seems to fix my usecase.

Ubiquiti Unifi Access Points just work.

I too have bought a variety of high end consumer devices (for a small business), and they all crash or randomly drop out.

Disable wifi radio on router, plug in Unifi AP, solid connection.

They are annoying to configure, but have been so worth it.

Apple AirPort has also been rock solid.

Motion seconded on Unifi. I've just deployed these and they're so far a delight. The PoE (power-over-ethernet) support is great -- if you can get the ethernet cable to the AP's location, you're good. Allows for zero visible cable installations if you have the opportunity to do in-wall cabling.

I did spring for a "Unifi Cloud Key", which is just a small (pack of cards sized), dedicated host for running the Unifi Controller. Makes administration and updates a fair bit easier.

As someone else mentioned, they do require a wired ethernet connection, but the point of that is these are intended as (entry-level) enterprise hardware. Unifi vs. Amplifi is for contexts where 1) the administration overhead isn't a burden and 2) it's preferable to run cable and leave wifi bandwidth available for clients and/or the rich multiple network support (vs. mesh networking). Obviously, #1 means it's not intended for the general consumer market.

I wish they didn't do 24V PoE. I know they are coming from a WISP background, but a AP seems like something that should support 802.11af. It's not like it's something exotic.

Note that the Unifi AP AC Pro (which I'm using) are 802.3af/at (PoE/PoE+). The AC LR and AC Lite are indeed passive 24v. So there's that option within the line at least.

I have a Ubuquiti ERL3 with 2 AirPorts and a TimeCapsule. One I got them placed around the house in good spots I had no problems. I'm halfway hoping the Apple stuff will die or be outmoded soon so I can replace them with Ubiquiti Unifi Access Points. I have a Ubiquiti EdgeSwitch too and one of their cameras. Everything I've gotten from them has been rock-solid. The configuration isn't on an iOS app but the forums/community are great.

My office just upgraded to an ERL3 and Ubiquiti Unifi on my recommendation and there hasn't been a peep about internet problems since.

Ubiquiti Unifi APs each require an Ethernet connection. They do not form a mesh network. That's the whole point of amplifi, and this new Google wifi product

Latest AC APs can't, but previous models all supported mesh networking.

> I wish I could easily add new hotspots that just seamlessly worked so that I could eliminate the dead spots on the edges of my house and in my yard over time.

Ubiquiti has wifi APs and you can add them whenever youd like. They automatically detect the existing APs and download their config.

I could be lighting a fire by saying this, but I've been using AirPorts for 5 years and haven't had a single issue.

Is the issue just that the non-Apple-customer market didn't have a good analogue?

> I've been using AirPorts for 5 years and haven't had a single issue.

And I've had Airports that just dropped things--all the time, randomly.

The Airports were a very mixed bag and would go from being excellent to piles of dogsh*t and back again with software upgrades let alone hardware cycles.

I use a Belkin with no issues. If WiFi drop outs were such a major issue all of the major vendors would have released products years ago.

This very much has the feeling of a "marketing" product designed to trick people into thinking it will solve internet speeds. When really the issue is the ADSL connection.

If you are using a Belkin and you aren't constantly monitoring it for security updates then it is behind on the security updates. That is the problem with mass market wifi gear. Before I got my OnHub I was either upgrading the firmware on my Asus, or power cycling it because it had crashed again, at least monthly. I've had my OnHub for about a year and I haven't even seen it once since I installed it.

Disclaimer: I am employed by Google but I do not work nor have any affiliation with this product, this is just my opinion and not that of my employer.

All the experience I've always had with wifi in the last ~10 years has been abysmal. I play games and make a lot of use of internet (downloading, working through ssh sessions, etc) and the slight hiccup and packet loss is very evident in my system and frustrates me a lot. When I am playing games and there's a slight interference, I notice immediately. When I am chatting on irc through ssh and my keystrokes don't go through due to spotty wifi, I notice immediately. Sometimes it's unbearable. I think this idea is really great. I had to finetune my wifi so many times in all the apartments I've lived at. Using channel analysers to pick the correct one every once in a while (due to recurring congestion/interferences), living in a very crowded area I often have ~30 access points broadcasting and cluttering and it's really obvious in the performance of my network.

In the last apartment I was staying at (moved out a few days ago), every time my flatmate at the opposite side of the house would open the living room door to walk to the bathroom, my signal would experience a latency spike of 100-200ms (average would be 10-20). We had a pretty poor router, to be honest, and I was at the other side of the apartment, but it was simply ridiculous. At times it would just refuse to work and I couldn't browse the internet for a few minutes simply because of too much congestion.

I think this idea with multiple smart access points is simply genius and while I plan to use wired connection for my next apartment, I'm definitely looking forward to a better implementation of the currently abysmal wifi that we have in general.

but wifi extenders have been on the market for a long time and are easy to use

The number one "family tech support" question I get these days is "why doesn't my internet work?" Usually meaning it's too slow, drops out, or doesn't work in certain parts of the house. Often they've done something like adding a cheap "wireless repeater" expecting it to fix the problem, and just making it worse.

Anything that allows me to just say "go buy this" instead of having to troubleshoot their cheap ISP-provided combo access point/router is a win in my book....

In my NYC apartment, the physical WAN connection is in a rather unfortunate location. There's no good spot to place a router in the center of the apartment, and no unobtrusive way to run cable there if there was such a spot. As a result I have to choose between a wifi dead zone in my bedroom, or a strategically located repeater. I've chosen the latter.

If I lived in a less dense area, in a wood-framed house of the same size, I have no doubt my wifi would be just fine. But with all the interference from neighbors and all the reinforced concrete, my home wifi definitely has a problem that requires something more than the simplest, cheapest router to solve.

This Google product might work very well for my use case, though I'm happy enough with my current solution.

Unless you have lead walls, a ~1000 sq ft apartment should be fully covered by a single access point.

The problem in NYC is that there's just too much band crowding; adding more access points is only going to provide an improvement (if any at all) until everyone else does it, too, and then it's going to be worse for everyone.

We really need better solutions for extremely high-density deployment of competing WiFi networks.

Unfortunately, it doesn't take lead to attenuate WiFi signals. Many turn-of-the-century buildings used chickenwire under plaster on the interior walls. This makes a pretty effective Faraday cage:


My SF apartment is the same -- well under 1,000sqft, need an enterprise router and a substantial repeater to get signal on the end opposite from the cable modem.

It's not just old buildings, either – at a previous employer we learned that a lot of the interior walls had a similar mesh inside. You could put a Cisco AP set to the highest power levels on one side, move a laptop one foot away on the other side and barely be online. We solved that by adding extra APs, which wasn't a big deal since we had the budget and plenty of Ethernet drops but in a residential building - especially a rental - that'd have been an annoying amount to pay.

My house, built <20 years ago (UK), has something similar. Currently at 4 APs (DD-WRT routers acting as APs, all hardwired to a central router), and still some rooms only get a patchy signal (fortunately we're more than 50ft on all sides from other building, so not too concerning about interference). 5GHz is even worse - for any practical use, it's pretty much limited to a single room per router.

Going 5Ghz is definitely a step in the right direction. Not only does it provide greater bandwidth for more channels but the reduced range actually helps avoid interference.

The downside is that you might need more access points to achieve equivalent coverage. Forming a mesh, like these Google APs do, is useful for avoiding additional wiring.

The beam-forming features present in 802.11n and 802.11ac also help reduce interference.

  The problem in NYC is that there's just too much band crowding
802.11ad will help there: runs on 60GHz. Certainly won't penetrate drywall. https://en.wikipedia.org/wiki/IEEE_802.11ad

The walls may not be a primary problem, but they do contribute. The simple test (which I've done) is to stand either side of a wall and check the signal. With an unobstructed line of sight to the router the signal is acceptable; on the other side of the wall the signal is poor (dropped ping packets, high latency, etc.). I doubt it's because of the additional 1 foot of distance from the router.

But to your point, I'm sure it's all interrelated and if I had concrete walls with no neighbors it would be significantly better.

I used to have wifi issues, but then I bought one of the UniFi Long Range wifi models.

I really can't imagine a house that would not be entirely serviced by a single one of those.

I still get 3 bars on my iPhone 30 meters down my driveway, let alone in the house.

I got one because we use them at work, and I was still getting usable internet in a supermarket 100 meters from my office across a parking lot.

They are designed for commercial use so initial setup is a bit complicated and requires java, but I have to recommend it.

Getting range in open air is easy (relatively speaking), but obstructions (walls etc.) can have a huge effect. I've managed >100m on a WRT-54G (with a slightly increase output power and a 7dBi antenna); same setup wouldn't reliably connect between my study and living room (~15m distance) due, presumably, to the metal in the walls.

UniFi LRs are really great, however I found it hard to get hold of the 5GHz ones.

UAP-AC-Lite and UAP-ACs and UAP-AC-LRs are available next-day-delivery from Amazon... I recently bought a UAP-AC-LR (to compare against my OnHub) and had it two days later via Prime.

They now even allow basic AP config via an Android/iOS app, so I didn't have to install the Java-based controller on a Windows box...

Apartments can have this issue - where I live there are 50+ access points visible and even with the router directly next to our streaming device I cannot stream Netflix HD (75 mbps plan with Comcast).

edit Not saying that Google's system fixes this but some users require nicer routers for that connection. I should note my actual computer gets an ethernet because I can never achieve on wireless what I can on wired!

This wouldn't solve that problem. Switching to Ethernet or finding a free channel (good luck) is the only solution.

Deploying software support for dynamic transmit power control would help a ton without requiring more wired infrastructure or more spectrum, both of which tend to be extremely expensive. When your laptop is in the same room as the access point, there's no need to be transmitting loud enough to clobber your neighbor's traffic. See https://github.com/thuehn/Minstrel-Blues

Not the only solution technically, put up grounded metal wall panels or wire mesh on the walls. Mounting tape/nail them on the walls if you live in an apartment.


Well it might. Google centrally manages each OnHub to actively pick the best channel, where "best" takes into account nearby networks and their channels and other spectral problems.

> Serious question: whose WiFi is not working, to the degree that they think, "I really need to get a modern router to make this Internet thing actually work."

With stock ISP router, I see occasional connectivity problems with several devices connected and one of them is streaming video (even though there is more than enough upstream bandwidth capacity to handle multiple devices streaming), all reasonably close to the router, and greater problems when one is farther from the router (even without any of the streaming video), in a relatively modest suburban house (<1500sf).

Anecdotally, I've seen lots of similar complaints from other people.

It's not about the internet not working in general. These routers are about dead spots in places that have functional internet already. It's a pretty common problem that absolutely does exist.

I live in downtown Manhattan and have that problem, but like OP asked, I don't think it's a huge enough problem to warrant a separate router just for the purpose. Most people simply use the router rented out from their internet provider (for free), and in most cases just make it work somehow (instead of buying a new device just for routing they would rather acknowledge that wifi doesn't work well in some parts of the apartment and live with it). Sure sometimes it's shitty and I have to reboot it to get it to work again, but I can live with it if it's not too often.

> Most people simply use the router rented out from their internet provider (for free), and in most cases just make it work somehow (instead of buying a new device just for routing they would rather acknowledge that wifi doesn't work well in some parts of the apartment and live with it). Sure sometimes it's shitty and I have to reboot it to get it to work again, but I can live with it if it's not too often.

OTOH, for lots of people the cost of a new router is a small fraction of even a years ISP service, so if it significantly improves performance of the internet service you are paying for monthly, and is usable for several years, its worth it.

Do any ISP really provide for free a router? Every place I've ever known will sneak it in as a couple dollars a month if you don't provide you own.

They often come free if you sign a longer contract.

Especially at the level of a $300-$500 investment, I'll add.

Sure, I might upgrade for $50-100 if it's more reliable.

A few people in larger buildings are indeed having an issue with connecting to the base point.

This access point 'network' uses each other to make a mesh network, increasing the chance that the nearest access point has a connection to the base access point, which has a connection to the internet.


I had wireless issues in an urban area (Downtown Oakland) where every router I tried had 10% packet loss in the 2.4Ghz band (5Ghz band was fine, but less coverage).

We got an Eero to try to fix the coverage problem (I lived with several housemates), but it didn't let us choose between 2.4Ghz and 5Ghz, it would switch dynamically. Certain IoT devices we had were 2.4Ghz and worked intermittently. No matter what router we used, no matter how expensive, the issue there never went away. There were ways to force the 5Ghz band on the client side using some BSSID hacks, but it was cumbersome. So we had to switch back to a single router setup with two networks, one 2.4Ghz and one 5Ghz, and urge people to only use the 2.4Ghz network if absolutely necessary.

I live elsewhere now and use a cheap 802.11ac TP-Link router without any issues in a house for a 200mbps connection.

The 2.4 going out was probably whenever one of your roommates or neighbors used the microwave. :)

Just talk to anyone who uses BT in the UK. Their HomeHub is atrocious. I replaced mine with a Draytek modem and high end Asus router and not only have my speeds gone up, I now get emails from no-IP asking me to manually renew my dynamic DNS domains because the connection routinely stays up for months (as opposed to days) at a time.

Edit: Oops, I should probably say something about improved wifi as well, as that was the point of the parent comment! :P

Needless to say my ISP's aforementioned modem router struggled to cover our typically average (for the UK) 3 bed semi. I even had a wired network point installed in my office upstairs, which I've barely used since we chucked it out. We can pick up the Asus router half way down the street!

Yep, I'm on BT with a 200mb fiber connection, and I use to use the default HomeHub. It was failing daily, the speed would drop to like 3mb, and you had to restart the router to get a proper speed again, connection would not always work. Just like you I bought a high end ASUS modem (120£, way more that I ever paid for a modem) but it was totally worth it. The wifi connection is now always top of the range, it has never needed reconnection, it changes everything. Definitely don't regret.

the number of networks i see living in a densely populated city (NYC) is pretty intense. Most modernish routers indiscriminately broadcast on both the 5GHZ + 2.4GHZ band, plus create an automatic (or, at least, one-click) guest network. So in a given building there maybe something like 2+ SSIDs per apartment.

that alone creates plenty of noise to wade through...

Do multiple SSIDs from a single AP create more interference?

I'm in Chicago, in a dense-but-not-a-high-rise area. LTE on my phone was drastically better than wifi on any device -- on weekday evenings, my laptop was useless. In a 700 square foot apartment. Bought an OnHub, now shit works. Google is definitely not full of shit when they say that the routers are really solid at optimizing the setup.

I could've fussed with placement of the old router, hunted through the channels, installed different firmware, or whatever -- but I don't spend that much time at home, and that's no longer what I'm interested in fiddling with.

depends how large your house is, the material of the walls etc.

I personally do need multiple routers to get wifi throughout the house.

I have a Netgear router and it's fine. Sometimes I have to reboot it but it seems related to Apple devices that refuse to give up their prior IP address after they've been away, causing conflicts when they return and the address has been assigned to another device in the meantime.

I'll raise my hand to this. And I'm not a mere mortal, I've run wifi for 1000+ attendee computer conferences. And I have problems with my home WiFi to the point that I got a OnHub a year ago to try to help.

Oddly, it basically did solve it except in a couple places in my house until a few weeks ago when the drop-outs got a lot worse (again, at the edges). The conspiracy theroists will love that. :-)

A real problem is that 2.4GHz is crowded not just by wifi but by other things. My old AP I couldn't use when I used my Microwave, and it is a brand new high end microwave. Not something I'd expect to leak. The OnHub did much better than the previous Buffalo (which in other locations has worked great for me).

So, you want to go 5.2GHz, because there is more spectrum. But there is also less interference because --- wait for it --- it doesn't go as far and penetrate walls as well. Which reduces interference. But it really means you want an AP directly in the rooms where you use WiFi the most. Which means you want something like this where multiple APs can cooperate.

The OnHub also is supposed to have stupendous antennas and multiple radios, even allowing it to do spectrum surveys. That's according to the marketing literature.

But it's not just me. My inlaws just built a 4600 ft^2 house and their A/V guy installed a AP in the centrally located laundry room. Wifi was terrible in the living room, and unusable in half the other rooms. Not sure if it was a crappy AP, or the location or what. I put another AP in the Livingroom and solved most of their pressing problems. But when I go back at Christmas I expect to put in an OnHub or Eero or AmpliFi... They have Ethernet all over the house, so I want something that can do wired meshing like the Eero. The AmpliFi doesn't seem to support it, IIRC, and the OnHub doesn't say anything about it with todays release, but the old OnHub is supposed to support it, according to a line item in their FAQ. I'm skeptical until I see it.

My brother in laws new house ALSO has similar problems. He had wireless drop outs, and was going to get the Eero, but I pointed him at the AmpliFi, which he got and has been VERY happy with.

If Google WiFi supports wired meshing, I will probably get that. If not, I'll either get a second OnHub, or maybe switch to the AmpliFi, especially if AmpliFi supports wired meshing and PoE. Ubiquity's cameras support PoE, so I'm hopeful there.

So, do these fancy products solve a problem that doesn't exist? Well, I'm glad your wifi works so well. But I'm definitely having problems at home, and am not alone. At least at work the WiFi is pretty uncomplicated, oddly enough since we have steel stud construction.

Me. I'm in the market for a new wireless router. My N66U drops out hourly. It never had great coverage. I have tried different firmware in an effort to save it but I believe it's dead.

I had the same reaction.

After some fussing, I bite the bullet and went back to Xfinity. I have a Netgear AC1750 router and an Aaris modem and I've never had coverage or speed issues in my house.

Any house with double pained impact glass, for Hurricanes can severely limit your Wifi signal.

This isn't about Internet. It's about streaming video.

It's about slowly caching the top 100 YouTube videos and then moving them quickly to whatever device you want to watch them on.

Despite their claims of being designed for user privacy, I'm wary of anything Google puts out. What information does it report back to the mothership?

I feel the same way. Was also disappointed with my Chromecast after it just stopped being able to connect to wifi and eventually wouldn't even boot at all. What kind of warranty and privacy disclosures does this solution come with?

I'd guess everything that can conceivably be reported.

I'd suggest unless Google's providing a fully open source solution, you should never trust them enough to purchase a networking device from them.

I'm surprised nobody here has yet questioned the wisdom of allowing Google access to literally all of your internet traffic. This is actually a commercial product that you pay money for, so maybe they're all aboveboard here, but with Google's history of trying to get as much personal data as possible about everybody makes me unwilling to trust them with something like this.

On a different note, is there any reason why someone should prefer this to Ubiquiti's AmpliFi? Unlike Google, Ubiquiti has a long history of making networking gear. The only obvious benefit I see right now is Google Wifi starts at $129 for one unit whereas AmpliFi starts at $199, but that $199 includes 2 "mesh points" and presumably to get the same effect with Google Wifi you'd have to shell out $299 for the 3-pack.

I realize it's popular to peddle conspiracy theories about Google in some circles but think about just how many people use Google search, Gmail, Chrome, Android, etc. – do you really think their choice of routers matters more?

There's also an inconvenient truth that such problems are greatly hindered by … Google in their prominent role improving security for average users. If their goal was to spy on people, they'd be the last to push HTTPS, HSTS, HKP, various TLS hardening, etc. rather than one of the first.

> conspiracy theories

Conspiracies do happen from time to time. Unfortunately the term "conspiracy theory" is often used to refer to unreasonable theories that are more about reinforcing preconceptions about what ought to be instead of observing what the situation actually is.

Obviously it is important to check any theory, and extraordinary claims require extraordinary evidence. In this case, however, the claim that Google wants to get their hands on as much data as possible isn't an extraordinary claim. It isn't a conspiracy which is usually defined as a secret plan. Google is very open about their business model based on analyzing large amounts of data with modern techniques.

Calling concerns about giving even more data to Google a "conspiracy theory" not only ignores Google's business model, it also dilutes the term making an actual secret conspiracy harder to identify in the future.

> think about just how many people use Google search, Gmail, Chrome, Android, etc.

The only service that is similar to sending all of your traffic to Google is GA (Google Analytics), which most people are not using willingly. Yes, a lot of damaging analysis can be done from search/mail/etc, which is why it's a terrible idea to also include the rest of your internet traffic.

> improving security for average users

Google has done a lot of work to keep user data safe from third parties. Very little of that work prevents Google itself from seeing your data. When the default configuration isn't end-to-end encrypted, in practice this means Google is still seeing everything. Google even calls this a feature that is part of their "assistant".

> If their goal was to spy on people

So you don't think their business is based on spying on users in various ways for better ad targeting? Even though Serge Brin and Larry Page personally explained their intention to spy to CA State Sen. Liz Figueroa?


None of those technologies you listed make it harder for Google to collect information from you, but instead make it harder for 3rd parties to intercept that data in transit.

If they were attempting to collect information via your router, those technologies would absolutely hamper them.

Exactly right

You mean in transit like through one of their wireless APs? :-D

> I'm surprised nobody here has yet questioned the wisdom of allowing Google access to literally all of your internet traffic.

This is what I think of every time I hear of Google Fiber.

I just bought an OnHub last year. WTF Google?!

And you just got an email saying that OnHub and the new Wifi router use the same OnHub firmware, which presumably will be maintained at least until the Google This-Time-We're-Serious router comes out.

Do you have a link to this message?

OnHub works with Google Wifi and will receive all of the updates announced today. https://support.google.com/onhub/answer/7168220

> I just bought an OnHub last year. WTF Google?!

Do you get upset when a new phone is introduced a year after you bought your last one? A new car?

The OnHub still has a disabled USB port and is missing other features - it's been abandoned.

The problem is Google sells products like the OnHub with indications that more is coming. It was strongly hinted that the OnHub was going to be the center of Google's soon to come Google Home solution.

Of course, those of us seasoned with the experience of buying Google products before should know better, but it really doesn't change the fact that people who bought it are absolutely right to feel upset.

This is part of Onhub, which can be a part of the mesh, and they both use the same app.

They released an 'updated' OnHub 3 months later; they really don't care about early adopters.

It was the same hardware, with a different headline feature. This _is_ part of the onhub ecosystem, and onhubs will be able to be a part of this mesh.

doesn't make me feel good buying this one knowing that next year they might abandon it....

Did they abandon the OnHub?

The above commenter was probably referencing the Nest situation

I'm a little concerned that an android device is listed as _required_... i would hope all functions of the wifi devices can be configured/administrated from the web or just an internal browser pointed to the device(s).

There are onhub apps for ios and Android


There is no Web interface; you really do have to use the app.

Well, that's a deal breaker. I don't own an Android device. There's no excuse for not providing a web interface.

I own an Android device and it's still a dealbreaker.

I thought it was cool to deploy a web application from my smartphone but after having done it once or twice, and two or three times having to hotfix production stuff from my smartphone, I can say it's not a device made to configure anything except alarm clocks.

In this case, OnHub has pretty much no configuration other than the SSID and password.

That's terrible. What about:

- Proxy settings - MAC address cloning - Port forwarding - Strict security controls - Wifi bands, 2.4 vs 5GHz (I guess this is less important) - Logs

It does have port forwarding buried under advanced settings but I didn't see the others.

iPhone or Android.

Okay, that's a bit better, but it should still have a web interface.

There are security implications: How do you provision a cert for it? Do you really want it running a full web server?

The native application is pretty great, because I can manage it from anywhere. (Which includes managing my parents one from anywhere.) I've honestly not needed a web interface, after getting used to it.

A web interface is reportedly on the roadmap, but not a high priority. The last time it came up on reddit, the question asked by the PM was "how would you use it differently than the native app?" And the nice thing is, he takes those suggestions and feedback seriously.

That's the one thing I hate about my Apple airport - great wifi, but I need a fscking app in order to configure it.

My Airports work just fine and have worked for years. I think that the WiFi issues are mostly that people are only willing to pay like $50 for their WiFi router. At this price point I'm not sure it changes anything.

If you try to use multiple Airport routers together without Ethernet you get severely reduced bandwidth. These routers (and other similar products) have multiple radios which allow mesh networking without reducing bandwidth as much. In a large house where running cable isn't an option having a mesh network can solve a lot of issues. In a small apartment a single access point is probably sufficient.

They have multiple radios and my bandwidth is basically capped by my comcast internet at around 180 Mbps which I get everywhere in the house with 2 airports, one extending the other.

I actually just got a "Wi-Fi booster"[1] for my place to have good Wi-Fi upstairs and it's amazing (it even has an ethernet port)

[1]: https://www.amazon.fr/Netgear-WN3000RP-200FRS-R%C3%A9p%C3%A9...

Anyone notice they are using their new TLD?

Yeah, I thought they were only using that internally. Well, there used to be an April Fool's page at elgoog.google which was just a mirror-image of the homepage, but that's gone now.

I think you mean http://com.google, which is indeed now just a redirect. I think someone discovered a vulnerability in it.

I guess the lay persons question at this point is why not just publish the software and dispel any notion of privacy infringement?

The list of features is missing one important item - Using these we will be better able to monitor your indoor whereabouts.

This is a nice piece of recent research: Keystroke Recognition Using WiFi Signals

[0]: https://www.sigmobile.org/mobicom/2015/papers/p90-aliA.pdf

And then what? What's the hypothetical secret evil plan?

> The system uses a technology called mesh Wi-Fi (something usually only seen in expensive commercial installations).

This is misleading. You can create your own mesh networks with several off-the-shelf routers / access points using open source software like batman-advanced [0].

[0] https://www.open-mesh.org/projects/batman-adv/wiki

I don't think its misleading to anyone -- the kind of people to whom aftermarket open source software for routers is a reasonable solution won't be misled, and its completely accurate for the people for whom what comes built in with the router is the only thing of interest.

But that's different from saying "something usually only seen in expensive commercial installations". And it is still misleading, accuracy of some information has nothing to do with what someone is looking for (a self-contained, plug-and-play mesh network).

> But that's different from saying "something usually only seen in expensive commercial installations".

No, its not. The fact that there are inexpensive alternative means by which it can be deployed, and by which certain segments of the population can and do deploy it, doesn't make it any less true that it is usually only seen in expensive commercial installations.

> And it is still misleading if it tends to mislead one group of people more than another group.

I don't think it tends to mislead any group at all. I think that the group to whom the exceptions to the "usual" case accurately described here is relevant will be well aware of it and so not misled, and the group who is not interested will also not be misled by the accurate statement, even though they are likely to be less aware of the nature of the alternative.

>> But that's different from saying "something usually only seen in expensive commercial installations".

> No, its not. The fact that there are inexpensive alternative means by which it can be deployed, and by which certain segments of the population can and do deploy it, doesn't make it any less true that it is usually only seen in expensive commercial installations.

Well then they should have made that more clear. Expensive + commercial means, well, expensive, and commercial installation means proprietary and hard to access (which, by the way, is also untrue).

> I don't think it tends to mislead any group at all. I think that the group to whom the exceptions to the "usual" case accurately described here is relevant will be well aware of it and so not misled, and the group who is not interested will also not be misled by the accurate statement, even though they are likely to be less aware of the nature of the alternative.

Sorry, I edited my comment for clarity. Anyway I don't think it's right to just say that because a certain population can't be mislead, it means that the information itself is not misleading.

> Well then they should have made that more clear.

"Usually" generally means that they are exceptions, but they are less common. There's no lack of clarity here, you are just ignoring that word.

Of course I am ignoring the word "Usually". It's not a substantial word. How much is usually? There are always "uncommon exceptions" to some subject (see the implicit flaw with this sentence? That's because I don't use the word "usually". What kind of an argument can I make without the word "usually"?)

It disregards a whole field of volunteered hard work to make mesh networking technology accessible by people who don't want to buy "expensive, commercial" hardware and software.'

It sends the message: "Don't look further, because it's not worth it -- because usually other solutions are expensive and commercial."

> Of course I am ignoring the word "Usually".

Then you are criticizing something other than their actual description.

> Then you are criticizing something other than their actual description.

No, I'm criticizing the wording. Perhaps you should ask yourself more questions about the construction of these descriptions and what messages they convey, than merely the definition of the word usually.

> No, I'm criticizing the wording.

Well, except that you are expressly ignoring a key word in doing that.

The problems with this are the same as the rest of the things announced today(except the pixel, in my opinion): Too expensive, and not much to set it apart.

I believe Google Home is cheaper than Amazon's Echo too. But in general I agree with you.

I've used my Airport Express to extend my Airport Base Station for years. Is this "mesh networking?" How is Google WiFi different?

If you ran an Ethernet cable you probably get a similar experience. Otherwise Airport routers use a single radio for communicating with the other access point and clients. This greatly reduces bandwidth.

Why are all these mesh Wi-Fi products (Google Wifi, Eero, UBNT AmpliFi, ...) only available in the US?

Ubiquity products are available in Europe.

"The AmpliFi online store is currently available only for US-based customers."

You have to buy from a European reseller.

Since the power plug is built into the repeaters, I assume I won't be able to plug it directly into my Australian wall socket.

Maybe. I've modded more than one device from the US for AU power by changing 1 capacitor for a higher voltage unit.

I believe European mains is 220 volts and our Australian mains is 230 volts, so it should be fine. I think.

Might need an adapter or change the plug.

Anyone know if this is going to make use of modern AQM (fq_codel or cake) like OnHub did? OnHub used the Qualcomm brand name for it: "Streamboost".

How is Mesh networking different from a wireless repeater? Yes, there's a nicer UI for this, but are there any technical advantages?

The difference between the two isn't in the UI. Wireless repeaters traditionally have a set receiving and transmission point. Let's assume that signal interference / reflection / amplification does no matter, for the sake of simplicity in this example. So if you have a chain of three repeaters A <--> B <--> C, and then you move access point C closer to A (closer than B is to A), then the topology does not change (do you see the issue here?). Worse, if you have a longer chain with a few alternatives, then a single point going down (call it X) would require nodes previously connected to X to be reconfigured.

Mesh networks automatically judge the path of least resistance to a destination, and distribute the state of all the access points on a network. If one access point goes down, all the access points previously connected to the faulty access point will find alternatives if there's anything else within signal range.

In this case, if those are meant to be portable access points, you can move them between different rooms without worrying about nodes choosing a suboptimal path.

Mesh-based Device Triangulation

You can tell this product is made by people who live in suburban McMansions. In a 1 BR NYC apartment, you have no problems with wifi coverage.

> You can tell this product is made by people who live in suburban McMansions. In a 1 BR NYC apartment, you have no problems with wifi coverage.

More relevantly, for instead of by (though its quite possible both are true), but, sure, in a small urban apartment, coverage isn't going to be your big issue, and you aren't going to need a set of routers forming a mesh network.

OTOH, lots of people don't live in small urban apartments and use the internet.

So they didn't build this product for your use case. Is that supposed to mean that they can't try to solve the problems that others have?

I live in a small house in SF, and I have wifi problems. Plaster and lathe sure adsorbs all that wifi signal.

Such a weird and negative comment. Are you alright? I mean, such negativity surely has got to come from somewhere?

Or, just not 1BR NYC apartments. So, the majority of Americans.

To be fair, if you live in a masonry building or even a particularly thick-walled reinforced concrete building, you might have trouble covering all of the rooms of a 2-3 bedroom apartment with a single AP.

Have you been to Silicon Valley? It may be a very wealthy area, but you wouldn't know it from the square footage counts of what people live in.

So split it with a couple of neighbors...

They do sell a 1-pack, although in that configuration really the only benefit is automatic security updates.

You won't have problems with wifi coverage. You will have problems with massive interference.

Introducing a new way to spy on you? A WiFi repeater will do the trick...

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact