Google Fiber is the fastest broadband internet provider in the US, but according to the Chocolate Factory's CFO Patrick Pichette, its 1Gbps connection speeds are going to increase ten-fold. Pichette told the Goldman Sachs Technology and Internet conference that the firm is actively developing the equipment to pump 10Gbps of data …
Four days ago, you published an article on Netflix's speed report in which Google fiber managed an average of 3.78Mbps download. My ADSL gets twice that speed, but is over Ten Times slower than Google fiber. They need to sort out their existing service before touting more rollox. The US consumer is being pumelled where it hurts in a severe manner, have they been completely brainwashed by the Stalin style capitajism in the land of th free?
I wonder what the ratio of users who use free Google Fiber at 5Mbps is to those who use the $70/month 1Gbps service. That may have something to do with the Netflix results, but I'm entirely speculating.
It also showed Google Fiber as the fastest network. Isn't it possible that the speed you refer to represents the maximum bandwidth for a mix of HD and SD broadcasts from Netflix?
They are, after all, a streaming service, not a download service. As such, they have nothing to gain from letting users download much faster than the program they are watching's maximum bit rate. Otherwise, when a viewer stops watching half way through a program, the bandwidth used was wasted.
The results at Speedtest.net suggest Google Fiber delivers a lot more than the Netflix survey suggests:
If I watch Netflix on my Wii on Google Fiber, it too will show slow speeds. Stream speed != Internet speed.
Something we can look forward to
Something we can look forward to here in the UK in say 200 years from now.
'Getting faster broadband into people's homes and businesses was vital to Google's own business, Pichette said. The more people used the internet, the better'
Ok but I have broadband, I have internet, it's there when I want it.
However this, 'Speed-Business-Productivity-Necessary for Business' equation is pure fallacy.
What is more important is getting everyone connected with a good and reasonable connection rather than claiming supersonic speeds are a must.
No one can read or type faster than a blink of an eye, so why don't they manage speeds to suit human beings. And as for 4K where is the market?
3.78 is just the average speed of actual streaming to Google Fiber users - obviously, Netflix won't have any content encoded to stream at 50 Mbps (full BluRay rate!) or anywhere close - probably some new HD/ultra-HD stuff at up to 10, plus a load of SD at nearer 1.
10 Gbps is just a pointless bragging exercise though: can you actually download *anything* at that rate, even from Google's own servers let alone anything the far side of a peering point? Can your computer handle that bitrate, given that most only have gigabit Ethernet cards?
Dialup to 512k cable modem was a huge step for me. That to 20 Mbps ADSL was an improvement; 100 Mbps would be a bit better still - but 10G? I'd never fill 1G anyway, even with 4K streams (BluRay peaks somewhere under 50 Mbps; even several 4K streams at once won't be 200 times that bitrate!)
Now, if I could just get a decent 100 Mbps connection - with fast upload, static IP, no "no servers/torrents/blah" rule, I'd be happy. Even faster would be nice, for the right price, but not a big deal.
It's all about reliability. Ten times the speed means many more ways to repair a missing byte. No need to resort to buffering (hold all the pixels) or changing to big pixels.
During a video conference, even a few big pixels can disrupt the confidence in reading bluffs by salesmen, field reps, etc. When each frame is magazine quality, suspicious body language can be picked apart in slow motion. Practiced liars will complain that their right to lie has been taken from them. So What!
I am a little confused by your response... I was there with you until you made the comment about reading speed and then discounting the benefits of higher speed due to 4k market.
First of all, at 270mb/s SD looks absolutely frigging awesome. Raw SDI signaling of SD video is truly amazing. I used to work a lot with film masters and often watched in original quality and it was insanely better. Higher frame rate is better. I recently saw raw 4K footage (which is heavily compressed since no one records at 12Gb/s) and raw 2k footage, both on 65" screens capable of 4k from 2 meters. Guess what? No difference.... Well until I looked at test signals which showed line art and hatch patterns.
4k is just not interesting outside of movie theaters.
So... 10Gb/s business case. To start with... Because we want to know if we can. For the moment, 4k is the only use case. At the rate which technology is evolving and the world is using it, real-time broadcast becomes less interesting except for sports and emergency information. So, we don't even need 10Gb/s for video.
Where can we use it?
- Offsite hard drive storage
Not so interesting since Google is also building Chrome which idealistically would run the app in the cloud too, so you only need bandwidth for remote desktop type of transfers.
- Remote gaming
Because running high frame rate games remotely is laggy, so eliminating latency issues with brute force instead of QoS makes sense. After all, QoS only works if you extend the trust boundary to the client which would be a nightmare.
- Video conferencing
Works pretty good at 384kbs in most cases, 9mbit for awesome quality, so even DSL should be fine. Most issues now are at the endpoint, not the network.
Honestly, I have no idea what the business case for 10Gb/s are. I personally don't notice the difference between when I'm on 1gbit at home or at a customer who has 40Gb/s access. The issue tends to be that the servers I connect to aren't fast enough to handle the load.
P.S. Downloading a film on iTunes from a PC with a 10Gbit NIC and a 40Gbit Internet connection rarely gives me more than 3Mbit rates.
If you're a business with perhaps 10 - 100 machines connected to your cloud facilities via 10Gb/s with VOIP, teleconferencing, online training, storage, backups all going via your Google Fibre connection then I'm sure Google can see the symbiotic (some would say parasitic) relationship as advantageous to both Google and the business using their services.
Re: Something we can look forward to
200 years? THAT soon?
"If you're a business with perhaps 10 - 100 machines connected to your cloud facilities via 10Gb/s with VOIP, teleconferencing, online training, storage, backups all going via your Google Fibre connection"
Then you're screwed... If everyone is VOIPing, teleconferencing, doing online training, ... who's left to actually do any work and pay the bills?
There are very few companies out there that can use 10Gbps. They'd almost all be on the server side of the cloud rather than the client side.
ie, the people sroring data onto the cloud only need 100Mbps or so, those serving the remote drives to thousands of customers will need 10Gbps... but those can be located in serverville, no need to have 10Gbps into every house.
That's the speed Google connects to Netflix, if all you need to stream HD is 3.78 Mbps then that's all it will record needing. Netflix doesn't do network speed test to max our users internet for obvious reasons.
Re: Carp @James
Whilst I don't disagree with your statement, I feel I must point out that as more people use things like Netflix et. al. this will result in more than one streamed feed into a household.
If the ultimate goal is to get TV services off of satellite and dedicated cable, then your average house may end up with three of four streamed services happening at the same time.
I somewhat hesitate to bring up my own household, because it is currently atypical (7 adults in the same house, all with their own media consumption devices), but even now, I can have a Sky download, a Netflix session, iPlayer/4oD/ITVPlayer/5 On demand sessions on connected BluRay devices or game consoles, and a NowTV session running live sport running similtaneously, together with gaming and YouTube. So you are not considering ~4Mb/s, it is multiples of this, and will become more if 4K TV over the Internet happens.
And I believe over time that other households will grow to resemble mine more.
Your point about 1Gb/s Ethernet cards is valid, but if the access box has 4 Ethernet ports, it could handle 4x1Gb/s. You are just thinking too small!
I would love to live in an area where I can get in excess of 10Mb/s. My exchange is not even on the list to be upgraded to fibre, and there is no cable TV provision either. It's just plain old ADSL 2+ annex M for me which has only ever delivered a maximum of 12Mb/s, and since a micro-filter failure is only showing a connected speed of 9782 Kb/s. As a result, when the scenario I paint above happens, I get some real arguments about who is hogging the bandwidth. Fortunately, my Smoothwall firewall tells me!
What's the point?
Name ONE THING you can do with 10Gbit that you can't do with 1Gbit. One thing that doesn't involve torrents, or dreamland technology like a personal holodeck served via the cloud.
For that matter, it is hard to see many things you need much more than 100Mbit for. I'm not saying "nobody will ever need more than a gigabit", but until we start doing something useful with hundreds of megabits it seems rather silly to worry about going from a gigabit to 10 gigabits. Especially when the userbase is measured in tens of thousands of customers, rather than the millions of us who live in the real world and won't see Google Fiber until 2040 at the rate they're moving. At this point, it seems more of a publicity stunt than anything else.
I'm sure large businesses would love to locate in areas where they're provided 10 Gbits, but I have a feeling the price for it will be a couple orders of magnitude higher for a business user, because there's a chance they might actually USE most of that 10 gigabit link and if customers actually use it Google can't provide it for anywhere near the prices they're charging.
Re: What's the point?
56k -> 2Mbps is a worthwhile step, so is 1->10Mbps. But 10Gbps??? WTF for?
As a kernel developer form way back (2001 or so), I started out with 56k dial up (all that was available). Start a download, go mow the lawn (half an acre with a push mover) and when that was finished the download was close. So when I got the option of 2Mbps I jumped at it.
But once you can download a movie faster than it takes to make the popcorn (100Mbps or so), there's very little advantage in a x10, x100 or x1000 speed up.
Re: What's the point?
Well what's relevant today is the uplink speed. And that's the current bottleneck. And there the difference between 1 and 20 MBit is still huge. With 1 MBit you can barely stream SD, with 20 MBit you can stream decent HDTV. Particularly with fast mobile networks getting affordable, having access to your local media is a big deal.
There is a second point. We now still have companies like Google who have the money to invest in the research needed for the next decades. Who knows how long we can still do that kind of research.
We also need to invest in the infrastructure. Our current last mile is still based on copper, a technology which became outdated in the mid 1970s. We desperately need to start a programe replacing it with a dedicated pair of fibres into every home.
Re: What's the point?
'Name ONE THING you can do with 10Gbit that you can't do with 1Gbit. '
Stream 10x as much 4K multi angle 3D porn.
Re: What's the point?
> it is hard to see many things you need much more than 100Mbit for.
Reliable 100MBit is sufficient for me to watch netflix in HD while my girlfriend surfs and watches youtube on the same connection. 100Mbit means a raw download speed of 10 megabyte/second, sufficient to download a triple-A game on Steam in less than 5 minutes, so... yeah, I don't know anything I'd really need 1Gbit or higher for.
100 mbit is fast enough that most problems are due to server or network issues (e.g. Google's horrid youtube cache servers), not that I've maxed the connection.
Maintaining network neutrality is IMO far more important than bandwidths higher than 100mbit.
I agree that the uplink speed is a problem in asymmetric situations like typical DSL and cable modems. But Google fiber was already 1Gbit symmetric, so that argument doesn't apply here.
Re: What's the point?
One thing you can do is host a huge datacenter with enormous capacity in your basement. Of course the TOS prevents you from hosting even a simple streaming server for personal use, so it is f*ing useless and is plainly a marketing stunt.
Re: What's the point?
> "Name ONE THING you can do with 10Gbit that you can't do with 1Gbit. One thing that doesn't involve torrents, or dreamland technology like a personal holodeck served via the cloud."
Run a large site (when large is >1000 users) without ever having network problems related to bandwidth.
For home users, your absolutely right; it's pointless.
Re: What's the point?
On a single fiber you can have 300Mbps down and 300Mbps up. The last test I did I received 315Mbps down and 270Mbps up. Later this year the service gets upgraded to 1Gbps in each direction and once again, it is just a single fiber, not a pair. How does a single fiber work? One wavelength down and a different one for up.
As for the outdated copper. Many houses are not setup with good copper inside the house and the owner is responsible for that. On the telco side, to replace all that copper will cost a lot of money and require equipment on a curb somewhere. No one wants to pay for the copper replacement nor have a box near them.
Re: What's the point?
How is fiber to the home a solution for bad copper INSIDE the home? If you're drilling a hole in the house to run fiber inside it, it is easier and cheaper to run a cat5e/6a cable through that hole.
If I could visit my house in 50 years, I am sure its occupant will have faster internet than the 25M/5M VDSL2 I have, but I wouldn't be at all surprised if they are still using copper. They'll probably have replaced the thin RJ11 satin cable I ran from the telco service entrance of my house to where my DSL modem is located with something newer by then, I imagine :)
This is showboating, pure and simple. As a few others have said, the extra bandwidth is no use at the moment. I'm lucky enough to have FTTP 330/30mb in my house, we're reasonably heavy network users and my peak average over an hour is 60mb. Working for the ISP that provides the connection I know I'm not shaped and the bandwidth is there whenever I want it, I just can't use it!
PS, due to the delights of fibre and light, you only need a single fibre in each house, not a pair...
> This is showboating, pure and simple.
No, it is competition and putting some pressure on the ISP monopolies.
I don't see anyone under pressure? If Google were to invent some technology that could get 10Gb at an affordable price to millions without spending billions in cash first then there would be pressure.
Google aren't doing anything clever, certainly nothing that any other ISP or telco couldn't do. They don't have the cash to fibre the world, so they'll do a city or two and will pretend they're a big connectivity player to further their cause. Showboating.
==> PS, due to the delights of fibre and light, you only need a single fibre in each house, not a pair.
Telco's are as tight as a duck's ar*e but even they run two fibres in - in case of failure. Maybe not BT, but whoever claimed they even think ahead?
They'll even install a camera so you can wave to the NSA guys.
I think this is the media companies answer (well, part of it) to the piracy problem ... give people fat pipes, and let them stream hi-res movies and sports. This costs money from the supplier side, and the download sites will not be able to compete.
Yes I know there are holes in this argument and other ideas are welcome :-)
10Gb Ethernet Kit
Is eye wateringly expensive
It's not the hardware cost
Broadband needs to be revised. Planning, changing out a bunch of boxes and confirming that they all work costs maybe 10x more than the difference between a 10G-box and a 1G-box. If Google can make any home-office into a real office,why wouldn't it. If ISPs, telcos and cablecos want to provide more value at the same price, why wouldn't they?
I wish the Googleian overlords would hit up my city I'm paying $70 a month for 3Mbs...
Hell their FREE speeds is faster then I get at $70 a month from comcast... And yes I know Google is evil, but are they more evil then comcast, or AT&T(where I was paying a year and a half ago $80 a month for heavily capped 1Mbs...)?
Laugh no more telco
Google is out to take all the valued customers. For about the cost of buying a better thermostat maker, Next, Google could wire big city condos and suburban homes where professionals live. Cities full of less valuable business are free to ban better internet and protect ISP incumbents.
"Getting faster broadband into people's homes and businesses was vital to Google's own business, Pichette said. The more people used the internet, the better it is for Google, and he opined that cloud services would become much more popular if bandwidth lag were eliminated."
yes, so google can know everything that you do...
Maxing It Out: It's what we do
The available bandwidth will always be utilized to the full. Our uses expand to fill the available capabilities.
When dial-up was used, you did what could be done on dial-up.
When you moved to 2M bit broadband your usage changed to fill that available capacity.
If you have a 100M bit connection, your usage will fill that capability. Currently, people can think of what will fill that capacity. Think back to when you had dial-up, did you truly believe that you would be able to max out a 2M bit or 5M bit connection?
When 1G bit or 10G bit become available, your usage will fill that available capacity, perhaps with streaming content or immersive 3D words, or something entirely new, that is not possible without the huge available transfer rates.
Meanwhile, in Australia...
...we're being told by the f***wits in power that 25 mbit is "all we'll ever need", which is why they want to piss away billions on FTTN under the pretense of saving money that was being "wasted" on FTTH.
Re: Meanwhile, in Australia...
For most people 25Mbps is actually an improvement compared with the idiots who preceded them. How a government could possibly attempt to build a FTTP network and at the same time plan for 50% of connections to be 12Mbps and in 2028 hope that 1Gbps connections would be approaching 5%.
For those who can justify faster speeds, fibre on demand will be available and that should be significantly cheaper than Labor's 1Gbps plans.
The reality is our best hope in Australia would be for the government to get out of the way and invite Google to build the network.
...the current government basically went to the polls telling porkies by promising a minimum of 25 mbit via FTTN - as far as I know, not a single ISP on the planet makes such guarantees for FTTN. Furthermore, the projections NBNCo made in regards to service uptake were exactly that - projections, and early uptake figures demonstrated that their estimates were quite conservative.
And before you go around calling the previous government idiots, there are buffoons among the current mob who are claiming that fibre has been superseded by copper! I'll just let that sink in...
I could not saturate 1Gbps
Well, I could, but I have no reason to. If I could get this at a decent price I would take it though. I would find something useful to do with it. I don't think people would have a lot to use it for now, but the few who do, why not?
By the way, 4k resolution VP9 encoded video is only 15Mbps for Netflix. Presumably you could easily do 50 of those at a time on gigabit without too much trouble.
I think this is more Google shoving the cable companies' "people don't need gigabit" back at them.
Google as an ISP? Doesn't that mean TOR-only connection for people who have a slight idea of what privacy is? All that's left is Google becoming a cellphone network operator (whatever you call them) and the company that "safeguards" our medical and bank records.
It's a good thing
I dislike Google's business model for various reasons, but telecoms providers are one of the worst type of corporations on the face of this planet.
They "invest" in a bunch of lines, then are content to sit on them and charge extortionate fees (as much as they can get away with) to "connect" you to the world, always attempting to create a monopoly. They don't continue their investment into their infrastructures if they can help it, they don't play well with others, forming alliances to protect regional "turfs" and only tries to find ways to extort more from the consumers.
Now I know, almost every corporation does that sort of thing, but the thing is, having a "tech" company like Google getting into the mix sort of encourages the advancement of communications infrastructures, and I hope other "tech" companies would get into it and bring in that technology company mentality to continue R&D and improvments on the infrastructure.
Facebook etc who is trying to encourage more of the world to get on the internet really first needs to break the regional deadlocks which is the cost and quality of connectivity around the globe.
How many home users will have the hardware to actually make use of 1Gps let alone 10? Beside having a fast enough network card there are other bottlenecks such as hard disk write speeds. 20 machines at full throttle may be able to make use of such speed or specialised hardware which all point to this only being of use to business properties at this moment in time.
Getting homes fibre connected so that any speed can be offered to them in the future is not to be sniffed at though. Personally I think most home users do not need faster than a 30Mbit connection at the moment but I would jump all over a 1Gb/s connection if it became available. A want, rather than a need.
You will understand, soon enough.
Humans are a strange species but offer some interesting non-logical thinking of value to the collective. Once a being approaches immortality their respect for efficiency can wane, you humans have such fleeting spans your immediacy and constant evolution can resolve issues that affect other terrestrials and transitory lifeforms.
One drawback is the maintenance of your current bodies, these are seen as archaic and an impediment to flexibility.
Once sufficiently embraced by his Googly appendages you will become greater than the sum of your parts.
And yes as will all things, there are two sides, look now upon the flying spaghetti monster and do you not also see his nemesis the antipasta!
The choice is yours bandwidth now or empty promises.
Make the connection!
(and mention my user name when you do, I get a small commission)
America is a technological disgrace (but good at PR bullsh*t)
First in colo(u)r TV, first in cell service, first with TouchTone dialling, first with the atomic bomb, first with stealth, etc BUT LAST IN INTERNET.
There are MANY places in the US which don't, even today, have wire telephone service STILL.
Of course being first doesn't mean you have the best - look at NTSC (Never The Same Colour) TV. GSM has always been technologically advanced (except for disabling encryption).
They can't blame bodies/square mile for poor service, Canada has areas with even lower body counts yet they have wired InterNet. Look at China - they have way more Fibre Optic than the USA.
Of course, developing/Third World countries have an advantage - we get the technology later but it is better. Korea has the best InterNet - it is simply eye-popping to see how fast new pages load. Japan has always been leading edge in equipment - they have bump-and-pay or dial-and-pay for years. My 'summer' cottage, way out in the boon-docks (sticks) in VietNam has 20Mbyte fibre optic service (to the pole-mounted DSLAM). We even had dial-up digital InterNet - digital into the house, that is, on every telephone drop in the country (It is removed when Fibre arrives).
Google is showing IT CAN BE DONE. When the rich, monopolistic AT&T gets competition it can move fast - even back in the day, it could move fast. When it wanted to. (TouchTone was promoted as it increased dialling speeds)
So now Google has proven it can be done, let all the cable and telco monopolies extract their collective digits and get with the times.
640k ought to be enough
This program won't be here past the end of the millennium so let's save a few bits on the date.
In the future there may be a telephone in *every* city.
If you build it they will come.
Google Fiber? On a cold day in hell...
Thanks to the fact that Google effectively owns all major websites via it's puppet industry (SEO), it's already hard enough to do anything on the internet without them tracking and profiling me in some way. Putting a data line into my house that they control so that every single bit that flows in or out of my systems can be tracked, monitored, profiled and sold to the highest bidder? That's just crazy.
No doubt Google will find some way to offer this for free once it goes wide-scale, thus tempting the uneducated masses, just as they have done for Gmail, Youtube, etc.
Not in my house. Not on my watch.
Better Make Asia Internet Faster
I hope Google also looking at More Bigger opportunity: to make Asia Internet Faster. By Making 100Gbps or 1Tbps World Network across Continent. Not all Internet User are from US. But also from Asia (now Asia is growing very fast (China, India and Indonesia) This is big opportunity
Google's position on net neutrality, with a contract that doesn't even allow me to enable a web server on my client PC, and serve personal web pages, profitless, from it on google's network, is appalling. Google, like the rest of the ISP oligarchs (there used to be 6,500 Internet Service Providers in the US, there are now, 10?) is trying to make us into passive consumers of content, not content providers. Sergey - first, do no evil.
Tell me then about your wife.
How beautiful she was, how tender and how true. Tell me with what speed she swung in the air, with what velocity she came off the wicket, whether she was responsive to finger spin, whether you could bowl a shooter with her, or an offbreak with a legbreak action. In other words, did she google?
Harold Pinter, No Man's Land, a Play by Harold Pinter, New York, Grove Press, Inc. 1975 p.32 [copyright 1975 by H. Pinter Limited]
- Review Apple takes blade to 13-inch MacBook Pro with Retina display
- Munich considers dumping Linux for ... GULP ... Windows!
- Game Theory The agony and ecstasy of SteamOS: WHERE ARE MY GAMES?
- Intel's Raspberry Pi rival Galileo can now run Windows
- Microsoft and HTC are M8s again: New One mobe sports WinPhone