Coming soon: Google Fiber 2 Gig
fiber.google.com
fiber.google.com
Unfortunately I moved out of a city metro area and am now in a smaller town that acts like a suburb of the major city I live near (~20-30 mins from the center of downtown). So my hopes for having google pay attention to us is ~0, at least over the next decade or two.
Spectrum offers 940mbps down / 40 up and it's... fine. I watch my bill like a hawk because of the history of all these actors and I really wish I had more upload bandwidth (for linux iso torrents - and I'm actually not being euphemistic here. I host a lot of stuff locally that I wish had a bigger upload pipe - passion project stuff in addition to seeding linux-related content.)
We're producing more for the world to see and upload is becoming increasingly as important for stable internet.
It costs something around $7/mo for around 3TB of transfer. That's three times the bandwidth I get with Comcast without paying an overage charge.
I also watch my bill like a hawk and I run a Speedtest exporter with Prometheus and grafana to make sure I’m semi getting what I pay for.
https://www.usatoday.com/story/tech/columnist/2020/10/03/att...
This leaves me with zero options for wired broadband.
I currently have gigabit service through comcast/xfinity. Tons of stability issues, but that aside even in the best of times it's as you say.. super low upload bandwidth. And even worse they still impose a terabyte bandwidth cap on me unless I pay them an additional $30/mo.
Well, I looked into the t-mobile internet. It's an LTE router, no bandwidth cap, a flat $50/mo total... and the upload bandwidth I get is around 40-60Mb/s. Download bandwidth between 60-100Mb/s.
Google Fiber arrived, and now Spectrum has "940 Mbps Internet", AT&T Fiber has "Internet 1000", and Grande (regional cable company) has "Gig Internet".
And in neighborhoods where they overlap, some of those providers have prices (introductory offers?) that are ~$10/month cheaper than in other neighborhoods.
In my area (a suburban area near Raleigh) it seemed like AT&T really got moving and deployed Fiber after Google put out the "coming soon" flyers. I hope Google keeps pushing forward, even if it's very slow. This will force other providers to improve.
I love maxing out my connection so would probably pay for 2gbps if offered but even I think its kind of a silly number at the moment. There isn't much consumer grade multigig networking equipment and the gear that does exist is easily 5x the price of the gigabit version.
Not complaining though! Any internet speed upgrade is great news.
Every time I turn on my PS4 there's it seems like there's some multi-gigabyte update that I have to apply first. I'd love for that to be pulled down in a few seconds rather than ten minutes.
I have gigabit and can barely max it out even when I start five torrents at once. By the time it's found enough peers, it's already done.
It'll be interesting to see what speeds we'll get on the PS5, where they are constantly advertising just how fast the storage is.
On Steam I consistently get 900 Mb/s, but I have pretty fast storage on my PC.
Eric Schmidt liked to say that more revenue is the solution to all problems at a company (give or take). I think Fiber is proof that that isn't strictly true.
High capex projects like this demand fiscal discipline. They demand the ability to plan, cost and estimate how long things will take. Fiscal discipline just isn't in Google's DNA. This is unsurprising given how insanely profitable AdWords is.
I also describe something like running an ISP as a hyperlocal national business. Think of these issues:
- Can you string up lines on poles?
- If so, who owns the poles?
- How do you get permissions to string things up?
- How can competitors delay and hinder that process (believe me, they do)?
- Do you dig trenches?
- What's the soil composition? I can tell you that digging trenches doesn't work well when there's a ton of small limestone rocks in the soil.
- Do you employ some form of micro/nano-trenching?
- Etc
And all of these issues vary from region to region, city to city and even within a city. The entrenched players have become very good at navigating this process, impeding competitors and capturing regulators and legislators through exclusive agreements, lobbying, protective legislation and so forth.
How Congress, the DOJ and the FTC continues to completely ignore what a shit show broadband is in the US is beyond me.
Revenue is still a solution, but you have to use it differently. People who just see all the technical problems their current ISP has assume that running an ISP is a technical problem; if you dig a little further (pun intended), it looks like an infrastructure problem, but that's not the root of it. Google Fiber is proof that running an ISP is fundamentally a political problem: how do you work around the many many things all trying to stop or kill you? (Hypothesis for one reason why ISPs are mostly terrible: the ones that exist were not optimized for good technology or good customer service, they were optimized for surviving politics, or built by companies who already had.)
There was a broader spectrum of decent ISPs when modems were cutting-edge, because then you "just" need to run a dial-in service. There were at least some decent ISPs when DSL was cutting-edge, because in theory you can buy your service from anyone. But now that fiber is cutting-edge, there's no solution anymore, because fiber wasn't treated as infrastructure from day 1.
There's a major reason why many "technology first" companies look into wireless and satellite to reach people. Not because those provide fundamentally better experiences or technologies, but because they can dodge the vast majority of local politics.
In an ideal world, we'd have fiber to every home as infrastructure, connected to a local meet-me room, and you could contract any ISP you like to light up that fiber, with a tiny flat fee going to cover the infrastructure. And then we'd have a spectrum of ISPs again, from budget to full-service to tech-friendly.
Or were they just stopping further deployments, and this only applies to existing customers/covered areas?
Looks like it's still halted in all those cities.
So I guess Fiber is limited to where they were as of 2016. [edit: in terms of new metro locations. Google Fiber seems to be slowly expanding in metros where they were already existing as of 2016]
As recently as October 19th, there has been news of expansion. A neighborhood association Facebook page post from that date says Fiber will start shallow trenching in their neighborhood, and some service could be turned on as early as November.
I just checked my address in the Google Fiber website and it says that it's "coming soon" to my home. I wonder what that means. Did they stop digging up roads? Are they doing something different now? What's changed in the past 3 years that allows them to suddenly come back to the cities that they abandoned?
NOTE: Just to be clear, I'm not completely unhappy with Google as I have AT&T Gig Fiber thanks to Google Fiber scaring the incumbents into action.
My favourite thing about it in Switzerland is that a bunch of different ISPs share the infrastructure, so I can get that speed from ~5 providers off the top of my head (Salt, Sunrise, Swisscom, iWay and Solnet).
It's so weird that the US has most of the world's largest tech companies but its technical infrastucture is such a mess.
[1]: https://bahnhof.se/
Network-wise, the US is a collection of local monopolies. It's feudal. Whoever has lordship over an area can squeeze the users dry while providing piss poor customer service.
Pushing my download speed up above 900Mpbs is basically a novelty. The only thing I use that actually uses that speed in practice is my Gaming PC's Steam updates, everything else takes a herculean effort. Speed testing websites are basically reporting their own bandwidth restrictions to me.
Even with that bandwidth, torrents struggle to perform at those kinds of speeds, and everything else is pretty much going to max out around 100mbps.
I have no idea what I would actually do with 10x more bandwidth. I could theoretically fill an entire 16TB hard drive with stuff in one day, but there's not a compelling need for that, at least not for more than a single day.
Im guessing they’re doing the 10gbit because the other option is bad cable internet for more reasonable options.
I love fast internet and pay for 1gbps fiber but struggle to come up with scenarios where I would ever come close to saturating that connection in the next 5-10 years. 4k Netflix is 25mbps. Estimates for 8k bitrate say it'll be around 100mbps. Maybe 10 years from all my household members will be hosting their hard drives in the cloud or something and 10g will offer marginal advantages but I just don't see it right now.
As you suggest, 10 Gb/s would be fun, but the time saved in practice would be quite minimal. I guess being able to play a game after 2 minutes instead of 6 would be nice.
Then again, I remember back in the late 90's when I got my first broadband connection. 10 Mb/s Ethernet connector installed in the wall of my apartment, I was amazed at the speed. I remember being shocked at how I could download a CD image (a Linux distribution, most likely) in only 20 minutes. It felt like the future had arrived.
Now, those speeds would mean several seconds to download a regular web page. I guess once a capability becomes available, you'll find a way to make use of it. I suspect that's true for 10 Gb/s internet as well.
Can one make good use of gigabit Internet connectivity? Yes, definitely. I wouldn't want to go back. If that qualifies as "need", I don't know, but it sure is convenient.
Community broadband seems to be the only real way to go these days.
I tried the latest Ethernet over Power and it was abysmal. Wifi5 is good when close to the AP. At this point it's wiring for Cat6 in order to fully utilize my Gige from Comcast.
Cat3 doesn't meet the spec for GigE, but in short runs without a lot of crosstalk from dense wiring, it's likely to work and worth a try.
If you only have 2-pair wires in the wall, you can still likely run 100BaseTX, which is OK if not very exciting. At least you'd remove most of the jitter of wifi.
If you're in a rental, I dunno how the landlord would feel about you messing with the jacks, but it's undoable. If you were considering pulling new wire, you likely own.
I do own, so I don't mind running wires. Though I am being honest, back in 2003, I ran my own Ethernet in my apartment from the main room to my office.
I looked at thin ethernet to run under the carpet, but I am basically going from the main entrance (where the TV/Cable Modem reside) to the opposite end of the townhome, where my office is.
Prior to wires, I'm going to try the wifi6 mesh system from linksys. If I can get ~500-600mb/s to my office, that is a worth tradeoff.
- no conduit in the walls - no ethernet in the walls - coax to every room but it all terminates outside the house
which I'm sure meets most people's needs but was annoying for me. And it would've been so cheap to do it "right."
So I'm relegated to relying on the fact that I have very conveniently located unfinished half of basement and garage where I can do lots of cabling and just pop through the wall right at the end.
I just moved into a 100 year old house and my first home project was wiring the whole thing for ethernet, resulting in a funny situation where my house has gigabit ethernet cables running past the original knob and tube wiring in the attic.
$30 / mo if you byoe. Or a little less if you're already leasing their equipment.
The only time I've ever maxed it out is with five torrents running at once, and that only lasts a second or two, because then they're done. And we're not a light usage household. It's pretty common for us to have three or four streams all running at 1080 at the same time, plus other big downloads and videoconferencing. One main limiting factor is that I've never seen single service provider get close to gigabit speed (like updates from Apple or Sony for example). It has to be a swarm to get even close.
I'm not sure what I'd do with 2Gbps. The rest of the internet just isn't there to support fully utilizing that connection.
But generally, at 1Gbps, the speed bottleneck starts to move away from the connection between you and your ISP and towards your ISP's connection to the rest of the internet. At this point, most of the residential ISPs start to fall apart.
I don't want to be misleading though, claiming "full gigabit" might be a bit hyperbolic of me.
By "full gigabit", I don't mean "exactly 1Gbps", but I'm easily able to reach the same speed to an external network as I'm able to obtain within my ISP's network.
For example right now if I run a speedtest to my ISP's Speedtest.net server, I see 820Mbps. If I run an iperf3 test to my Feralhosting server, I see 820Mbps. If I run a iperf3 test to my Hetzner server, I see 894Mbps.
It's not technically right at the 1Gbps that I pay for but it's close enough that I'm generally happy.
Just as they did it in front of my house, Google announced they were no longer expanding their fiber network. Now there are three orange cable butts sticking up from the ground at each end of the block, lashed up to telephone poles but not connected to anything electrically.
But I'm glad the people a couple blocks over from me are getting 2Gbps service while I suffer with Spectrum cable internet. Oh, they recently refunded the $10 they collected a couple years (or more) ago.
Only a massive cultural change, driven from above, can give Google long-term vision and follow-through. And it won't be from Pichai.
Truly incredible to see the difference between a company for which internet access must be commoditized and existing rent-seeking monopolies.
Anyone has experienced upgrading to WiFi 6 and care to share?
The advantage for me is, I have multiple things using the internet, and if they all get 300Mbps it's pretty good.
Haven't had a chance to try a WiFi 6 to Wifi 6 connection yet.
Paired with a WRT32x router running OpenWRT, these are the speeds I get on my 1000/50* GPON fiber connection. They do tend to vary depending on what speedtest server I use and when. I have seen it max out at 950 megabits however
https://www.speedtest.net/result/10403429271.png
*1000/50 due to political reasons, despite being symmetric capable
>PING 192.168.31.1 (192.168.31.1) 56(84) bytes of data.
>64 bytes from 192.168.31.1: icmp_seq=1 ttl=64 time=1.69 ms
>64 bytes from 192.168.31.1: icmp_seq=2 ttl=64 time=2.16 ms
>64 bytes from 192.168.31.1: icmp_seq=3 ttl=64 time=1.54 ms
>64 bytes from 192.168.31.1: icmp_seq=4 ttl=64 time=1.88 ms
>64 bytes from 192.168.31.1: icmp_seq=5 ttl=64 time=4.79 ms
>64 bytes from 192.168.31.1: icmp_seq=6 ttl=64 time=2.10 ms
>64 bytes from 192.168.31.1: icmp_seq=7 ttl=64 time=2.04 ms
>64 bytes from 192.168.31.1: icmp_seq=8 ttl=64 time=1.76 ms
>64 bytes from 192.168.31.1: icmp_seq=9 ttl=64 time=1.91 ms
>64 bytes from 192.168.31.1: icmp_seq=10 ttl=64 time=1.84 ms
>64 bytes from 192.168.31.1: icmp_seq=11 ttl=64 time=1.64 ms
>64 bytes from 192.168.31.1: icmp_seq=12 ttl=64 time=1.71 ms
>64 bytes from 192.168.31.1: icmp_seq=13 ttl=64 time=2.54 ms
>64 bytes from 192.168.31.1: icmp_seq=14 ttl=64 time=1.60 ms
>64 bytes from 192.168.31.1: icmp_seq=15 ttl=64 time=1.62 ms
>64 bytes from 192.168.31.1: icmp_seq=16 ttl=64 time=1.56 ms
AFAIK the spikes are due to Windows' network discovery.
Dense areas or areas with poor line of sight won't be well served by 5G home service. Long distance 5G is only marginally faster than 4G, and neither is anywhere close to the bandwidth of fiber.
Nothing is going to beat widely deployed fiber in urban areas.