Many comments are around not seeing the point for connections so fast when you'll get throttled or simply won't notice.
I think that's kind of the point of this experiment (under GFiber Labs). Exploring the challenges around offering this, and discovering use cases for it. It's a silly argument now, but the 640K argument keeps popping up from people who are just fine living in "the current past".
Do I expect Google to roll this out globally? Hell no, they can't do products or support, but it's nice that they run this kind of experiments. Also, it seems that just pushing the industry forward is enough for them, their real revenue doesn't come from Fiber or Fi anyway.
Now, on what new things can be attempted with this? Someone mentioned that this was saturating a single SSD reads, which is interesting.
For some applications my SSD does almost only reads, and the writes are mostly to update content. My disk must be 99.9%+ idle.
Can then some nix-store/IPFS like system host all the applications I run? I'd be kind of cool to have every package virtually installed on every computer. Games fall on this category too, where people just keep them on SSDs for content to load quickly, and the writes to SSDs only happen when games are updated (or downloaded, which would be a no-op here).
Google Fiber started a deployment in my city (Louisville) in 2017 before eventually giving up [0] after ~2 years.
AT&T, on the other hand, has been steadily rolling out gigabit fiber across the city since 2016ish, and I've enjoyed their service that entire time. It's consistently stable, fast, and the price is only slightly higher than Google Fiber's - I think I'm paying $90/month right now for symmetric gigabit.
The local Spectrum cable ISP guys offer ~300mbit (with a pathetic upload rate), and a Spectrum sales guy once asked me if I "really needed a full gigabit, no one actually needs that much speed." Lol.
I'll bite - I've yet to see a big difference between 300mbps and 1GBps speeds. It's completely marginal.
Being able to save 8 minutes downloading Fortnite isn't a big enough deal. For video calling, you're limited by how much bitrate the provider is willing to allocate for you. For looking at photos on social media, you're stuck with Instagram compression. Internet browsing more tied to the number of requests and client-side parsing speed than total bandwidth.
Should our society be spending hundreds of millions of dollars on laying new fiber for marginal improvements?
It's not remotely marginal when you have ~4 TVs at a time streaming 4K, two people working at home on video calls, constant cloud backup and so forth. I currently have 5Gbit, and will upgrade to 10 the second it becomes available. With better bandwidth _typically_ also comes better latency, and more competently managed DNS, though your own DNS setup is still likely better than whatever you privacy-invading-telemetry-hungry ISP sets up.
Curious - Are you really maxing out that 5gbit connection where you would need to upgrade to 10gb? I have the 1 gbps symmetrical service from AT&T and have never maxed it out. Even with 2 adults working from home and 2 teenagers. All of us usually streaming something on different devices on weekends at max resolution, plus game downloads via steam/xbox gamepass, backup jobs pulled from friend's homelab, my own backup transfers to another offsite location, in addition to me self-hosting bunch of stuff that I share with family & friends, I've yet to max out that 1 gbps pipe.
The fastest transfers I've seen from source is from Xbox game pass downloads on a PC coming it at 150-200 mbps. Even 4K streaming is usually no more than 20-25 mbps/stream. AT&T keeps sending out flyers enticing me to upgrade to the new 2gbps or 5 gbps service. But again, even during heavy usage I have plenty of bandwidth leftover. So I'm wondering how one maxes out a 5 gbps connection.
No, he isn’t. We do that with a 400Mbps connection at home. The hospital system I admin the network for doesn’t even see that, and we move a lot of data.
Yeah not just in bandwidth, but maintenance, distance, resistance to environmental factors -- I've seen fiber cans caked with mud and flooded, and still pushing data just fine -- are all massive improvements.
Plus, upgrading the optics on both ends can often lead to massive improvements without changing the lines, and on a scale that far exceeds copper.
Not when the source you're pulling from has no problem sending you bits at 4Gbps (accounting for overhead) like GCP or AWS (they have to be on like 1Tbps now or something insane like that)
Agreed. I think lag / ping is just as important as throughput in determining subjective snappiness of a connection (unless you're doing large up/downloads.
Agree, personally I don't see any reason for more than 100 Mb/s.
Teams limits screen sharing to 3-5fps, at 1080p that's 0.4 Mb/s according to their stats. Yeah, gigabit won't help my sufferings at all.
AT&T wasn't delivering 300mbps before rolling out fiber. They were selling 100mbps connections with data caps for more than what the cable company was offering.
Comcast is rolling out Docsis 4.0. I’d hardly call that Stone Age
OTOH, I have a rental where most people in the building have opted to cancel Google Fiber in favor of Comcast (even tho Comcast is more expensive) because Google Fiber has had multiple outages each year.
It’s still Stone Age, sorry. Their business is squeezing the most out of old cable infrastructure not building next generation infrastructure to support our increasingly critical digital future.
Comcast is rolling Docsys 4.0 out to support, “300 Mbps, 500 Mbps, 1-gig all the way up to 2-gig” service. Their words, not mine.
Fiber easily supports 10-gig, or 100-gig with easy upgrades. And here’s the point: that single strand of fiber can support, without replacement and in a single strand, around 44 Tbps. That's not something cable companies like Comcast want you to know. Even if they squeeze 10 or 20 gig out of this aging cable lines, it’s holding society back.
People don’t need it. But a T-1 (1.544 mbps) line was the recommended size for 45 people in 1999. Now not 25 years later that’s laughable.
> I actually thought Google had given up on Google Fiber years ago due to incumbents using regulatory pressure to keep them out.
I think they said they were going to focus on their existing markets, but not on expanding to new markets.
Which, to be fair, could be interpreted as a sign that they would shut it down at some point (this is Google, after all), but I don't think they ever said anything to that effect, and it was several years ago, so maybe they really are just focusing on improving service without expanding. Stranger things have happened.
> maybe they really are just focusing on improving service without expanding
I don't get it. It's a product the whole country wants, rolled out to only 25 cities and they're done growing? Can anyone make a case for why this isn't a horrible business decision (if true)?
I've heard it said that one of the worst challenges at Google is that nothing moves the needle in their revenue stream other than ads. They may have tons of other things they are up to, in huge demand everywhere, but why commit & work at anything that doesn't move the needle?
Having a foot in the ISP game is a good existential hedge. It gives Google practice at doing something absolutely required for their business model to keep working. If ISPs started doing dastardly things against the dumb pipe/network neutrality model or started raising prices astronomically, Google could scale up their ISP model to keep people connected to keep the ad stream going.
I do think it's a bad decision not to keep rolling out. But I think Google has a ton of difficulty maintaining sufficient caring for non-new non-shiny things, and is plagued by ads being so overwhelmingly bigger than everything else.
> I don't get it. It's a product the whole country wants, rolled out to only 25 cities and they're done growing? Can anyone make a case for why this isn't a horrible business decision (if true)?
The regulatory and logistical challenges to rolling out in new cities vary extremely widely, and this isn't their core business or core competency. Compare that to telcos, who focus exclusively on these problems, and it's relatively expensive for Google to compete in this area compared to telcos.
Of course, that's assuming telcos are investing at all (which they aren't, at least not when Google or municipalities aren't threatening them with competition), so this argument isn't automatically categorically correct. But that's the gist of it, and the truth is probably somewhere in the middle.
Selfishly, I want Google to expand Google Fiber, but I would actually rather have municipalities force existing telcos to deliver on the investments that they've already paid for (with taxpayer money, in many cases). At the end of the day, that's the real problem. Google Fiber is just a stopgap.
The skipped my burb 10 years ago because they were required to run underground. Last year, after they said they were giving up, they decided to go ahead and dig the whole town.
Clearly, they reversed course (again), and didn't announce it.
Even if I had a bigger pipe, Steam is the only service which I think might be able to saturate it. Everything else seems more limited by providers. Looking at you GOG, where I seem to be lucky to pull down 5mbps.
Supposing they offered this at my house, what could I do with it that I already can't do with gigabit fiber?
Google Fiber already offers 2Gbps to a friend of mine, but he hasn't upgraded because he doesn't see the point. Most of his bulk transfers (Steam, etc.) seem to be throttled at the source, and upgrading from Gigabit to 2.5GbE or 10GbE gear isn't cheap.
I could see this being useful for an office building or an apartment building, but what would a single family home do with it?
I know there's much more to real-world performance, but ha, that's slightly faster than the benchmarked top sequential read speed for my main (gen3) SSD.
Well, that's one possibility, and seems boring, but imagine everyone else's filesystem being available to you at almost local throughput (and some extra latency, but hopefully low).
I wish they also talked about latency, because now what I want is for cloud gaming, video calls, and other real time use cases to get extremely responsive and latency free. Beyond 1 gig, I'm not sure I'd notice much difference in my household.
For some reason, consumer Internet access has always been sold based on bandwidth - it's very rare to see any consumer ISP make any statements about latency, even for mobile providers (where latency is much more noticeable and problematic).
As far as customer marketing is concerned "speed" equals "bandwidth", even though there's a whole other axis that really needs to be considered.
It’s annoying that it’s near impossible to get a grip on that before signing up.
We live somewhere which has open access GPON. Been with 3 different providers, the average ping to cloudflare, google etc with one was 7, then 5 and now 2.5ms.
All are pretty good, but 2.5ms is nuts for a residential line. They should be selling based on that.
I'm not sure if this pushes competitors or not. Apparently nothing is done with the internet speed at our area. I live in Anaheim, and I only have 2 options: AT&T (50Mbps max), and Spectrum (100Mbps+). The up speed is still very slow, the down speed is always horrible for both, somewhere around 10Mbps. I can't do multiple video calls at the same time. It's been like this for 10+ years.
Recently I have Tmobile 5G home option and it's a bit better with the up speed, 20-40Mbps. The download speed is flaky, not consistent, somewhere from 50Mbps-250Mbps.
There has never been any fiber option, I wish this exists here, even with 1Gbps option is like heaven.
This is a frustration for me as well. Here in Boise, I've got CenturyLink 80/40, and that was a sweetheart deal gained by going through our CenturyLink Rep at work (my neighbors get half that). Technically I can get "up to a gig" from Sparklight, but with no IPv6, the same upload speed, a data cap, and 4-5x the cost.
I doubt very much, GFiber has been around here long in those cities (limited area) too, but for some reason, it has never tried to expand into other cities nearby.
I think Google's goal is to try to make the market more competitive, it doesn't try to take over other ISPs.
This is amazing bandwidth. Do the terms of service forbid using home connections for running businesses: mini data centers, remote storage, video serving, proxy point, etc.?
6 definitely isn't rare. For example, iPhones started supporting Wi-Fi 6 with the iPhone 11 Pro in 2019.
But isn't that how it always is? Whatever next version is going to be rare until it's not...
You have to start releasing products using it at some point.
Wi-Fi 7 is schedule to be released May 2024, but there have been devices released already supporting it for awhile now via the Draft standard. Just do a search for wifi 7 on Amazon.
The Pixel 8 supports Wi-Fi 7 for what it's worth.
I assume the iPhone 16 in fall 2024 will support Wi-Fi 7.
> I assume the iPhone 16 in fall 2024 will support Wi-Fi 7.
I wouldn't bet on it; Apple tends to lag behind on WiFi standards. The first phone to support WiFi 6E was this year's iPhone 15 Pro (the non-Pro 15 still doesn't).
"Additionally, by offering this service with a custom pre-certification Wi-Fi 7 router (another first ever — in fact, Wi-Fi 7 is not even fully certified yet) [...]"
Wifi is generally backwards compatible, so that just means whatever ISP modemrouter that Google Fiber provides is capable of providing Wifi 7 for newer/future devices and 6 and earlier for older ones.
Just a couple of decades ago, I started with 64kbps ISDN uplink on copper twisted pair and 2Mbps sat downlink, serving 100 computers in a college campus browsing center. That was clearly web 1.0.
And a decade ago, I was doing pairs of redundant 10Gig links to a datacenter.
That was web2.0.
For the past 5 years, I have a GPON network to my home (incidentally with a Nokia ONT). Today, this is the only way Internet is delivered to homes around me. But I'm on a 300Mbps plan and don't think I utilize anything close to it. There are 1Gbps plans available. Btw, my mobile already clocks 400Mbps+ in speed tests on 5G network. This is the post-covid WFH plus multiple 4K streaming services era.
I wonder what new use-case (like video streaming) will emerge that will push utilization beyond 1Gbps. Is it AR/VR? Is it streaming updates to big fat AI models running inside your home? Or is it streaming lots of data from your home to AI models in the cloud and getting responses back?
> I wonder what new use-case (like video streaming) will emerge that will push utilization beyond 1Gbps.
AR/VR are subject to latency. You'll be able to download the 2-3 digit gig files for the worlds quickly, though!
Serving files isn't _quite_ so latency sensitive, though. I can see "backup EVERYTHING to your home NAS AND our cloud!" being a more common scenario though. To be clear, we already have this but it's usually a single appliance limited to less than 10 disks/100TB and there is no "offsite" mirror; when you remotely access a file on your synology/wd...etc appliance, you're pulling it from the NAS on your LAN over your ISP.
Doesn't it seem like you've got enough at some point? I've never heard of someone demanding larger water pipes to their house or bringing in more electrical lines for higher capacity. For me, I've not really needed more than 100Mbit symmetrical though 1G certainly seems nice.
> For me, I've not really needed more than 100Mbit symmetrical though 1G certainly seems nice.
I think this is correct for "most" people most of the time. I'm less certain that'll hold in a decade or so.
Every once in a while, I need to pull down a big data set and it's nice to only wait tens of seconds instead instead of having enough time to go make a quick cup of coffee while I wait.
I have friends that are content-creator adjacent and they routinely sling around massive amounts of footage for editing and other production reasons. Depending on the video, target audience/platform and a few other variables, waiting for the raw footage or even the final export to upload and become available for distribution can cost real money. E.G.: if the press embargo expires for $theHotNewTech at 11 AM and you're still waiting for footage to upload, your competitors will have a video ready to go and YouTube will show their video and not yours. Being "first" on that platform for certain events has a MASSIVE impact on how well the video "performs".
I just fail to see why we are going above 1 or 2 Gig, considering even 1 Gig is fairly uncommon for people to have.
The fastest speeds I have ever realistically seen is on my Xbox capping out at about 650, maybe 700. I think steam caps out at about 150.
I don't say this as "how could you need that much speed" but more, realistically is anything going to be able to even return those speeds especially taking into account the bottlenecks between you and those servers.
Edit: Maybe if you have several Xbox's but even so, with 2 heavy gamers we have never had to have a situation where the other was impacted by what one was doing.
Is there any 20G infrastructure available for the prosumer? I've got 10G connections between my a few devices at home, but most everything is capped at a 1G connection.
QSFP+ cards are pretty cheap and deliver 40gbps. I used a pair of them point-to-point between my workstation and NAS for awhile, and it was pretty great.
The switches available for "prosumer money" used to be eco-terror brodozers and super loud and fussy. I think the new mikrotiks are interesting, and if I wanted to do 40 again, I'd probably go that route with some ConnectX-3 cards.
> We’re starting in Kansas City, North Carolina’s Triangle Region, Arizona, and Iowa.
Meanwhile I live <10km from their goddamn headquarters, am one of probably MANY people in the area with actual ideas for how to use 10 Gbps+, and I can't get more than 1.5 Gbps Crapcast. It makes zero sense.
Google dug up my neighborhood in summer 2023 for fiber and are now sending me ads every week. I've had AT&T 1 gigabit fiber for the last 5 years. My parents have the same thing and I do my backups from here to there but the fastest I ever get is around 70 MBytes/s (560 Mbit/s)
Even though speedtest.net often reports 940 Mbit/s the fastest real world download I've ever seen is around 650 Mbit/s
I'm generally happy with my speeds and 20G or even 2G sounds great but without knowing what my bottleneck is I'll just stick with 1G.
Does anyone else feel a little weirded out about filling in a lot of private information into a random Google Form? I mean it is linked from a Google blog post but it still sets off alarms in my head.
Maybe now I can self host all my wp sites from home. That would be fun /s
I have the Gogole og wifi box still, they acknowledged it didn't work well, replaced it with an Airport Extreme which is still working. I wish Apple still made the Airports...
I have 500M service and still never even get close to my bandwidth limits. It's upload speeds that are inconsistent and generally slow. Is this symmetric? Even if it is, what is the case for Wi-Fi 7 in the home, let alone 20G down?
I wish they’d bring faster than 1Gbps speeds to the Webpass network but I doubt that’ll happen soon due to the architecture (Ethernet to units instead of fiber).
Also worth noting that map is by county, but pale pink can mean only a town or two in the county has it available. If anything it overstates availability, despite how little coverage it shows.
I’m also bitter about it living in CT where we have almost no competition for ISPs so offerings are slim.
This is still just a PR stunt. I was on the waitlist for google fiber for SEVEN YEARS and when it finally got installed, what do you know, it never actually delivered more than 250Mbps.
I think other ISPs just assume the additional bandwidth is not needed for the majority of users out there. They’ll have to act once Google lowers price on some of the slower tiers
Same reason they originally created Chrome - they needed a browser that was faster and enabled more complex web apps, and they were sicks of the offerings at the time, so they built their own.
Absolutely. Google even announced Google Fiber was coming to some cities.... then the incumbents stepped up and offered affordable Fiber connections... then years later Google retracted the promise it was coming, and now will not offer fiber in those cities.
So basically they called their bluff, and that was enough.
I hope they stay in it because I love my 1 Gbps Google Fiber service. Install was great, it's super reliable (compared to Spectrum cable Internet), and the price is nice. I don't even own a computer or switch with network ports that can do > 1 Gbps, but 20 to the home is impressive.
It looks like it became a pet project. If they really took the goal of nudging the ISP industry towards better service through competition, they would care more about expanding their coverage.
I think that's kind of the point of this experiment (under GFiber Labs). Exploring the challenges around offering this, and discovering use cases for it. It's a silly argument now, but the 640K argument keeps popping up from people who are just fine living in "the current past".
Do I expect Google to roll this out globally? Hell no, they can't do products or support, but it's nice that they run this kind of experiments. Also, it seems that just pushing the industry forward is enough for them, their real revenue doesn't come from Fiber or Fi anyway.