Think big with a gig: Our experimental fiber network | Hacker News Reader

Think big with a gig: Our experimental fiber network

googleblog.blogspot.com

269 pointsashishbharthi149 comments

Time for the other so-called broadband providers to shake in their boots.

I like companies that have no respect for established business models. In this perspective, despite its size, I think it's fair to call Google a start-up.

I think this is more around trying to force the broadband providers to get their act together. Scare them ;0)

Right. At Google's scale, it will even benefit them a lot if other broadband providers provide a better service, since it promotes more use of the internet, and thus more ad income for Google.

Or will make video ads more practical. Uhm, yay?

I guess people said the same thing when pictures were first being used on the net.

I'd say that it was still a net benefit not to be text only, wouldn't you agree? Ads can be annoying, but they also pay for a lot of what we're using online every day.

The way Google throws products on the market to increase competition for the benefit of its core products ... it's genius.

Are there any other companies that do this?

Yes, tons. See "commoditize your complements". Microsoft is a master of this as well.

http://www.joelonsoftware.com/articles/StrategyLetterV.html

I don't know of any real market changers from microsoft, save IE, which was an attempt to kill competition and preserve market. Maybe I am too far from the microsoft circle to know of any...

Wasn't the XBox hardware pretty much sold at cost to sell their games for their own platform? How many recent games does Microsoft have for the PS or Nintendo?

All consoles (at least at launch, and for years afterwards) are sold at a loss, with profits made up on the games. The difference for Microsoft is that the XBox division has never made a profit, unlike Sony's or Nintendo's.

Note: the Wii is the only recent exception to this rule that I know of; the hardware has been sold at a profit since day one.

If I recall correctly the GameCube, DS and GBA were all profitable from day one as well. It's possible this goes back as far as the NES.

There is an important difference between compliments and complements.

Which is interesting, as Google certainly wants to commoditize operating systems. I'm sure Bing is intended to commoditize Google somehow, but I don't think it's working so far. Microsoft appears to be more in the role described for Sun in Joel's article, of the company trying so hard to sabotage potential competitors that they act against their own interest.

(Not nearly as spectacularly as Sun creating Java to spite Microsoft, and ending up commoditizing their hardware business. But I do not think Bing would look like such a good business strategy without the obsession of beating Google.)

Yes I know about that strategy ... but Google's means are different.

When they release a freebie (like Chrome or Android), they force their new competitors in that market to follow their lead.

The browser's speed, instability and lack of proper standards implementations hurts them? ... so they release a super optimized browser with a clear design and with tabs running in their own process. Now Firefox, Opera and Safari (at least) will follow.

The mobile-market is too closed and becoming more so? So they release Android, a pseudo-open OS with an online-store that has more relaxed rules. And I'm pretty sure they don't need an Android-monopoly (as Microsoft is trying with everything they release) to achieve their goals.

Their strategy is not (directly) for increased lock-in (as Microsoft does) ... it's more for eliminating distribution channels / middle-men that might get in their way on reaching customers.

They use the scare tactics a lot, but they didn't work in the wireless spectrum auction did they?

They ensured that open access provisions were included.

I forget where the link was (I think somewhere on yc), but it showed a chart of internet access speed through a few years, rated by countries. Most of the large-nation world had grown by 10%+, many over 50%, in a few years, while the US had dropped by a couple percent. This is in addition to most of those countries being higher already, so it's not merely a 5kbps -> 8kbps jump. My price for internet access not going down, and speed not going up (personal experience) in several years implies the ISPs are gouging because they can.

Proper competition is essential in every industry. If Google can provide it, good for them. This is an area where only the big players can compete, but nobody's stepped forward to fix things. Maybe this will be it.

Here's a chart: http://lifehacker.com/5451127/seems-like-everyones-getting-f...

Yup, that's it. While clicking around something like that site, I also found one that covered a couple years, and the trend is roughly the same (with a couple more spikes).

edit: another, related chart, from a quick Google: http://www.marketingcharts.com/?attachment_id=781

I'll grant that median is likely not a good measuring point, but it is still revealing. Hell, Canada is out-doing us, so you can hardly argue it's because we're more spread out.

Just as a side-point, although Canada has a large land area for a small population, it is actually much more urbanized than the US.

The problem is that cable companies were given monopolies by the local governments when the setup.

There is the possibility that the providers aren't intentionally gouging their customers. Even though the market is small, there is enough of a market to drive prices down. I think it's much more likely that those companies are just so horribly inefficient that it really does cost them a lot to provide you with high-speed access.

Several years ago my family got cable internet from Time Warner. The modem/router they provided was one of the first models to support DOCSIS 3.0. We stayed with Time Warner for at least 2 years, and they never rolled out DOCSIS 3 support at their end. The maximum speed we could get from them was 6Mbps. We're now using DSL, and the maximum we can get is 6Mbps, though with more reliability than from cable.

The nearest cable box is next to our neighbor's driveway. The phone company has fiber to their box in our other neighbor's yard.

There's enough competition to keep price inflation mostly at bay, but there clearly isn't enough to convince the ISPs to invest in the equipment that would be able to give us connections that are four times faster. Efficiency doesn't have a lot to do with it. For all practical purposes, our local ISPs aren't spending any money to improve service in areas that already have some service.

Non-growth (negative growth!) despite significantly more paying users and far more fiber lines than a few years ago? What have they been rolling all that fiber out for, if not to crank up the speed? Oh, wait. To sell the same speed to more people for the same price to make way more money.

I don't buy inefficiency. They're selling it at this speed for this cost because they can, just like cell phone companies. Which from a business standpoint makes perfect sense, but it's at the cost of screwing everyone else. Since significantly better businesses / strategies clearly exist if it's just a sign of business incompetence, why haven't the successful ones, say in Japan or Korea, migrated over here and decimated the market?

Average access speed is dropping because of the increased penetration of wireless access in the US lately.

Then what's the excuse for Japan being so much higher? They've had a lot of wireless access for quite a while, and web browsing on their phones for much longer. If that's taken into account, they should be significantly lower.

It's not an issue of wireless penetration, it's about maturity. Obviously the first rollout in the USA is going to be slower than a mature infrastructure like Japan's.

Surely when you roll out new wireless, you should roll out the fastest stable technology? Wouldn't Japan be held back to some extent by wireless infrastructure that hasn't yet been upgraded?

Yeah, after Google starts selling broadband significantly below cost, no real ISP will be able to meet customers' inflated expectations.

It's not in Google's best interests to reduce competition in the telecommunications industry. If they offer bandwidth at such a low cost that it threatens the viability of existing providers, they have to commit to providing that bandwidth in perpetuity. Their current business model requires bandwidth, and replacing that by raising bandwidth prices in the future when competitors have been eliminated would be met with government opposition fairly quickly.

I'm not worried about Google becoming a monopolist and raising prices; I'm concerned that they'll offer unsustainable (or even free) pricing in a small trial and many people will begin to expect other ISPs to match that pricing in nationwide deployments. Much like the complaints that US broadband sucks compared to South Korea, but worse.

I don't think Google is going to sell at below cost, the limitation on large network's bandwidth is primarily routing technology which has continued to improve while US broadband has stagnated. It's not like other countries have not been rolling out low cost 1Gbit consumer connections for a while.

PS: Despite all these talk of caps etc, upstream bandwidth is still a small chunk of a large ISP's budget.

Anecdotally: A friend of mine who lives in Sweden gets 100 megabits for about the equivalent of $25 per month. It's not a gigabit, but it's a lot closer to a gigabit than five megabits is.

There are some faster connections available in the US too. In and around the north east iirc.

Here on the west side of Canada, Shaw offers 100mbps connections for $150/mo. That's obviously a complete rip-off but it's nice to have the option to get gouged. They price the 25mbps connections at $100/mo presumably to make it seem economical to get 100mbps.

http://www.shaw.ca/en-ca/ProductsServices/Internet/Nitro/

Telus has 25mbps as well now. Fttn and capable of serving 100mbps according to a Telus technician I spoke with, so I'm sure they'll up the speeds at some point.

Nice to see competition that isn't just offering us "gifts" ... for signing 3 year contracts.

The variable cost of selling broadband is extremely low.

The problem isn't the variable cost; it's the $2,000 truck roll per customer.

$2000? citation needed.

http://www.ftthcouncil.org/en/knowledge-center/legislative-r...

Apparently it's down to ~$1,000 if you assume 100% take rate and don't attempt to serve rural areas, but it will easily exceed $2,000 in realistic deployments.

In Utah if you're willing to pay for your fiber upfront it's only $3,000. And then you still have to pay the monthly fee for Internet access on top of that. http://www.app-rising.com/2009/11/utopia_proving_new_option_...

Oh, I thought you were insinuating that any truck roll cost $2000. In this case, I can say that your original statement is incorrect: it's not $2000 per customer, it's $2000 per residence at most. (Consider college towns, where lots of residents cycle through few apartments.)

I'd be fine with a wireless last mile to avoid cable messiness.

Wireless gigabit? Does not compute.

Huge pipes to supply 600Mbit wireless stations (http://en.wikipedia.org/wiki/IEEE_802.11n-2009) would be a reasonable stopgap until faster wireless tech becomes available.

They wouldn't be selling "at a loss" based on fixed costs. They'll eventually recoup those somehow, much like AT&T or Comcast or anyone else will. Comcast may hope you'll buy On Demand movies while Google might settle for more Youtube views.

Is it selling at a loss if I pay someone $100k to write software and then sell it for $15/pop?

But at whose cost? Google already owns a significant chunk of their own infrastructure, so their costs are cheaper than most ISPs.

Google could profit at rates that would be "below cost" for other ISPs.

Isn't the poor coverage of decent internet connections a much bigger issue? Highspeed internet in urban areas might be cool but what about the rest of the country.

A bigger issue issue for who? Telecommunications companies have decided that it's not worth investing much in connections to low density areas. If there are enough people clicking on ads in those places to make it worthwhile, maybe Google will provide connections in those areas. If that doesn't happen, I don't see any justification for subsidizing people's connections above a level that is necessary for basic access of information to aid in making informed voting decisions since we share a government.

That's the thing, I think internet access is essential to being an educated citizen nowadays, and many websites are no longer viable for dialup users to visit. High speed internet for all should be placed on the same level of importance as education for all.

Satellite internet speeds are acceptable for web browsing and is available everywhere. I'd consider subsiding the costs of such connections for those who can't afford it. My understanding is that things like Universal Service Fees subsidize even those who can afford it.

Have you actually tried satellite service? Latency is worse than dial-up. We tried for a while to get some of our customers using satellite but that turned out to be a failure due to the miserable performance.

If you feel it is essential for the educated citizen today, then it isn't Google's job to address the lack of broadband in some areas, but rather the government's job. I'm fond of the idea of refreshing old programs like Rural Electrification by changing them into new broadband expansion programs. Whispers of this have been made by politicians including our President while he was campaigning, but as far as I know nothing even remotely substantial has happened to either promote or cancel the idea. Instead it just seems forgotten.

Sorry, yeah, this is what I was referring to, I wasn't trying to imply that Google is responsible for doing it.

Apparently, 63% of US adults live in households with broadband. Considering the number people who avoid or can't use the internet (i.e. older folks) this seems like remarkable penetration to me.

PDF: http://www.pewinternet.org/~/media/Files/Reports/2009/Home-B...

Plenty of old people can use the internet. I suspect more people don't have broadband because it's not available at a reasonable price for them than you might think. Adults living in rural America had home high-speed usage grow from 38% in 2008 to 46% in 2009. vs. Among adults whose highest level of educational attainment is a high school degree, broadband adoption grew from 40% in 2008 to 52% in 2009.

I've been wondering lately what would happen if someone like Google decided to build their own cellular network. Is there a way that someone could disrupt that industry?

Rumor Mill: Google Acquiring Sprint http://blog.tmcnet.com/blog/rich-tehrani/google/rumor-mill-g...

Thats from 2007. Sprint and Clearwire merged last year and Google is significantly invested in Clearwire's 4G network build-out.

http://en.wikipedia.org/wiki/Clearwire

Like what Gmail did to all the e-mail provider/administrators who said anything over a 100MB Inbox was too expensive.

This has nothing to do with technology. It's all 70's as all the rest.

it's all politics.

Expense isn't a problem when you have a multibillion dollar company who wrote their own filesystem to use custom designed free software on custom designed low power servers redundantly spanning over cheap disks in energy efficient custom datacenters and no uptime guarantees.

For most companies a new Exchange cluster backed by RAID SCSI disks and a backup system upgrade isn't a trivial cost.

Edit: yes, Google did it cheaper, but that's not because IT people were lying, misleading or dismissing - it really is expensive to do 'properly'.

Worth noting: We had something quite inspirational up in the Lakeland fells recently. One villiage (Alston Moor[1]) was so sick of constantly being set back by ISPs, that they organised a community action group to lay their own fibre optic cable during some planned roadworks. Everyone pitched in, cloudsourced infrastructure if you will, and with just enough people they look set to have 100Mb broadband before a lot of places.

The guys who got it off the ground set up www.cybermoor.org and are encouraging similar grass roots development for other remote areas. I like that.

[1:http://www.cybermoor.org/index.php?option=com_mtree&task...]

We need more of this sort of thing.

It's arguably far more important to get good internet access to rural areas than to inner cities. If you're in Central London or Manchester, you can quite easily shop, meet others in your field of work, network, be entertained without touching a computer if you don't want to.

Meanwhile in remote areas, it's not so easy to pop out and do these things and the internet has been a huge leap forward - but we're being held-back by spotty, slow, distinctly iffy internet connections that just aren't capable of sustaining meaningful economic activity. The advent of the internet is a huge chance to revive rural communities which have been in decline for years now.

But while BT, Google and other companies concentrate their high-speed efforts exclusively on profitable inner-city areas which already have a plethora of other options, we're in serious danger of leaving "forgotten" rural areas behind. I have friends on a Scottish island which currently only has 512kbps on its exchange...

Or downloading a high-definition, full-length feature film in less than five minutes.

Did you hear that? It sounded as if millions of hollywood executives suddenly cried out in terror and were suddenly silenced...

Why? They still control the content and the licensing. I don't think they really care about the carrier media (DVD or packets), they care about extracting every last possible dime from the consumer for the content.

If the content can be distributed AND billed effectively over this pipe, then it's fine. If the content can NOT be billed appropriately, it simply won't be licensed for this distribution channel.

I think their concern is that it will be distributed over this channel whether it's licensed for it or not.

The sad thing is that hollywood execs are scared by news like this when they should be looking at it as an opportunity.

The ability to download a full HD movie in <5 minutes opens up a treasure trove of revenue possibilities. It makes content delivery much more convenient and thus much more compelling for customers.

I think purchasing will become much more of an impulse decision— much like what the App Store has done with software.

Companies like sony have much capital tied up in movie theaters right? This would be a negative impact to those revenue generation mechanisms.

Was that what it was? Because I think it sounded a lot like scandinavia sighing that we've had this for a decade. (I'm sure other countries has too, the US really has backwater infrastructure.)

Admittedly this probably has a lot to do with luddite technology policy, but surely it has more to do with the U.S. having 10 times the land mass and 15 times the population as Scandanavia [1].

[1] Stats from Wikipedia, summing Norway + Sweden + Denmark area and most recent population vs. United States.

Actually, Norway and Sweden has a lower population density than the US has; we've got even more land mass per person.

http://en.wikipedia.org/wiki/List_of_countries_and_dependenc...

I think it has more to do with that we don't leave infrastructure development squarely in the hands of the market.

> Actually, Norway and Sweden has a lower population density

I realize that. The multipliers I gave necessarily imply exactly that: 15x pop. and 10x area => U.S. has 1.5x the population density.

I would imagine that non-uniformity of population (i.e., cities are high-density, backwater rural states are not) is exacerbated by the greater land area of the U.S.

Sweden and Norway has very non-uniformly populated areas as well. Don't blame your land mass, blame your telcos. They are screwing you over.

well for the most part, the US has no market for utilities - it is all monopolies

Government-issued monopolies. This is terrible stick to try to bludgeon markets with; there's hardly a market to be found.

If we've got 10 times the land mass and 15 times the population, and rolling out speedy broadband is more cost-effective given greater population density, doesn't that mean the US should be faster and cheaper than Scandinavia, not slower and more expensive?

Perhaps I'm not thinking about this clearly, but I'm figuring that greater population density makes it less cost-effective because you have more endpoints, and therefore need more network branching, more wires, more intermediate nodes, more digging, and more time per square km.

If your revenue is what you get paid every month by your subscribers, and your cost is your total mileage of fiber (plus switching equipment, upstream connection, etc.), then the more subscribers you have per mile of fiber in the ground, the higher your net income.

More saliently, the justification most often trotted out when US consumers complain about being so ill-served by broadband providers is the higher population density in Japan and South Korea.

Ah, I was stupidly neglecting the profits from subscribers after the infrastructure was in place. However, wouldn't this run counter to the OP's argument that it's the free market economics preventing this from happening? With the greater profits available afterwards, why wouldn't the money-grubbing bastards invest in it?

I'm not saying the reason for the US's poor infrastructure development is decidedly because the US has left it in the hands of the market, I'm just saying I have a feeling it has something to do with it, that it's one of the factors.

In Sweden, government policy and government subsidies have ensured widespread broadband coverage, not just in major cities, but in large but sparsely populated areas as well. This is true for both internet infrastructure as well as mobile networks. It hasn't been squarely up to the market to decide which areas deserve high capacity broadband and mobile infrastructure.

I remember back in 2001 when I traveled to the most northern part of the country to teach programming; Kiruna — the world's second biggest city by area, population 19,000 — had wi-fi network connectivity over the entire downtown area (which admittedly wasn't as large as it may sound). This was not due to the power of the open market.

In Sweden there is no such thing as lack of cellphone connectivity. Calls suddenly disconnecting does not happen. The only reason I even know of these phenomena is because I've heard Americans speak of it. And, finally, service is cheap, dirt cheap, by US standards.

Sorry to get meta now, but I think this is why downvoting is a bad idea. You're discouraging me (and others) from working through any confusion they might have.

Having a larger userbase should actually make your service cheaper since you can spread your fixed costs across more users and get serious economies of scale.

This certainly does qualify as disturbance in the Force :-)

On the other note, I do hope that Hollywood would embrace this as an opportunity not as a problem.

However, if anything in this life is certain, if history has taught us anything, it is that you can expect big media to cling to the old ways of doing business for as long as they can...

The repercussions of this on the next generation of gaming consoles would be very intriguing.

I thought the "download" part was funny too. 1g is as fast as many people's laptop hard drives anyway, so local storage is pretty much pointless.

Local storage still matters since bandwidth is not the only thing people care about: http://rescomp.stanford.edu/~cheshire/rants/Latency.html

Yeah, I know, I was sort of joking.

The hard drive certainly quickly becomes an OS-managed cache though like ram is for hard drives now though.

Don't have time to find the source, but latency is comrable as well.

This will shake up the broadband industry the same way ebook readers are transforming the publishing industry. This is HUGE.

I am in NY and I am waiting for years for fios in my area. Hopefully this adoption will move faster than FIOS.

It's as if all the collective anger towards comcast and time warner will now have an outlet.

You may have a problem in NY.

In general, the more built up a place is, the more vested interests there are, and the more they are able to manipulate the legislative process.

Google moves at Internet speed.

Out in the "real world", it's still illegal to sell cars and wine over the Internet, because the local car and liquor distributors have gotten their preferences written into the law.

You may find that Google can't go the last 20 feet to your house without buying off a few hundred interest groups, politicians, and unions.

Thanks man. I feel much better now.

Not to mention it's harder to build a network where everything is paved over and digging requires navigating through a couple hundred years of previous projects (and having to use union labor for it). I don't envy anyone having to plan a network buildout in NYC.

In many other parts of the country it's mostly digging through dirt and in when you're in a neighborhood using a water jet to tunnel under driveways. Piece of cake.

Rolling out fiber in NYC takes less digging than you might think, there is a lot of open space where huge bundles of copper wire where used by the telcos. It can become expensive to rewire a building, but cable companies did this not so long ago.

It's not just NY. It's a lot of places. Many towns have locked down franchise agreements for telephone and cable tv "utilities" just like they do for water or power.

If Google admits they're going to deliver advanced Google Voice or YouTube services over these lines, that's infringing on these agreements and the towns face lawsuits from the legacy carriers.

You can't just hop on poles and start stringing wires all over the place.

But I highly doubt Google is planning on offering "Cable Video." In the same way that VOIP is differentiated from Voice Service according to the FCC, offering IPTV does not require a Service Provider to register as a Cable Franchise. VOIP & IPTV are seen as "applications," not a native service type.

Google's own press release suggests they're looking for new ways to use ultra-highspeed broadband:

Or downloading a high-definition, full-length feature film in less than five minutes. Or collaborating with classmates around the world while watching live 3-D video of university lecture.

Now look at how Geneva, Illinois fought off AT&T U-Verse (called "Project Lightspeed" in the experimental phase) because they already had television franchise agreements in place and didn't want to get sued by the cable company:

http://www.geneva.il.us/att/Lightspeed.htm

(Note, also, that Geneva tried to set up their own broadband network and was roadblocked by AT&T. Karma is a bitch.)

I'm not suggesting that Google is looking to offer cable TV. I'm just implying that you can't waltz into a town and offer these kinds of services without raising the hackles of the existing companies that are providing utility services in most typical towns.

A town might willingly invite Google with open arms, but that doesn't mean it's a done deal by a long shot. Expect lawsuits.

Interesting, I hadn't heard of that. Of course there will be lawsuits, but my point was that by current definitions they should not be in violation of cable franchise agreements, even if they are allowing users to download full-feature films in < 5 minutes.

Therein lies the opportunity. An abundant inventory of dark fibre. The financial resources to lay the foundation for a new city. Everyone will have fibre to their door. I'm no fanboy, but Googleville anyone?

I suspect it will encourage incumbents to roll out FTTH only in neighborhoods where GoogleNet is available. For the other 99.9% of the US, nothing will change.

This may be paranoia, but I'm actually worried about this much bandwidth being the hands of ordinary people. My reasoning: super-botnets. This has the potential to very easily create a several 100Gbit DDOS network, which is essentially unstoppable. I understand the need and potential for high speed networks, but I think gigabit might be taking it a bit too far. It's a "with great power comes great responsibility" sort of thing. But perhaps I'm just being overly paranoid.

Don't most South Korean already have fiber to the home? The internet hasn't melted down yet.

I've never looked into it, or even thought about it until now, but I wonder what virus / bot infection rates look like in non-English speaking countries.

Most infections seem to result from an automatically created subject line in English... I'm sure there are viruses created specifically in other languages, but are they as successful?

If I am not wrong, this kind of speed already exists for few years in some parts of the world. S.Korea, Japan and England comes to my mind.

The maximum you can get at a reasonable price in England is 50mb download at around $80 / month. It's also worth bearing in mind that upload is the critical speed for DDOS and that is _severely_ limited on most connections.

The maximum you can get on a residental connection in the UK is around 2mb.

I live in Sweden. I had 10/10 mbit fibre in my (regular non-campus) apartment back in 2001, for peanuts, maybe $20 a month. In 2005 I briefly upgraded to 100/100 mbit (very common today), but found that it wasn't worth it (about $55) because what good will that do me when most of the internet can't keep up with it? I think it about doubled my speed from my previous connection, but that's it.

Today I have 12/4 mbit DSL, and I pay less than $20. I don't need more bandwidth, but if I did I could get a lot more than I could use for less than $50 a month.

Owning your net access, owning the DNS, routing through their backbone. These guys have all your bases covered.

How do you form an underground internet that bypasses the big players?

While I understand some of the concerns about privacy with Google, one thing I like about Google is that instead of shouldering competition through backdoor deals (like MS did with OLPC), they go out of their way with superior products and services.

They are late comers in search, gmail and Browser, but their success has been smashing thanks to superior products. You have to give them that.

http://store.cablesplususa.com/riinar.html

http://www.gardeners-world.net/prod_show.asp?id=853

Alternatively, look into http://en.wikipedia.org/wiki/AMPRNet

This is more than just kinda great, this is definitively awesomely great. Words can not describe how crappy Time Warner is in NYC.

Bonus points for doing this in a recession when traditional companies are running for the hills and laying off people left and right. Google continues to innovate and crush competition in multiple industries. I have no doubt that if they want to become an ISP or even pretend to become one to force traditional ISP's to innovate then good for them and better for us.

I think Google would be better off partnering with existing service providers for this project. It's a hugely complex process to start from scratch. Getting the pole rights alone to build infrastructure is going to take them years. In some area's within the span of a mile you might pass poles owned by half a dozen different entities. Most of them owned by incumbent Internet service providers who will tell you the pole is already at capacity and if you want to put up your fiber they need to replace 50 poles to cover one new street of infrastructure build out. In many places they'll have to dig trenches for new construction. In cold areas that means you lose at least 5 months a year of actual work time or you spend a fortune digging into the frozen ground. They're going to need to hire a small army of contractors to build plant and splice fiber or make the investment in man power, bucket trucks, gear, etc to do it in house. They're going to need warehouses and office space all over the country. And of course they'll need customer & technical support staff. I'd be shocked if Google makes any real progress on this outside of extremely limited deployments. The whole system is setup to make it very hard for any outside force to compete with incumbents.

This is very true, but google also has a vast amount of resources, and some of the smartest and technologically literate people in the world working for them. Its entirely possible that they have an army of robots or some game changing technology to do things differently from the norm.

Since it's a pilot project, Google can simply choose to install it in places that are easy to install. Of course, the results won't be representative.

Considering they are only looking for a couple communities at most (and maybe just one), it's pretty logical that they will choose one where it's relatively easy. Temperate climate, clear rights of way, city government assisting, etc. They're not going to choose Manhattan.

I watched Verizon install FiOS around my city several years back and it was all quite efficient. Contracting firms dug the ditches, another firm put the cables down while the digging crew came back behind and filled in the trench. It was amazing how quickly they could hook up a neighborhood.

I would think it would be good for Google to partner with the local non-telco / non-cable provider of utilities. The local electric or gas company (who is used to commodity business models) would be a ideal. They already have access to the houses and would make a good "franchise" partner. They probably wouldn't mind another "dumb pipe".

In israel the electric company is working on being an ISP. they use some kind of robot that lays fiber on top of electric cables.

I'm not sure this will have the effect on competition some around here think it will. The only way this will cause traditional providers to lower their rates and increase their capacity is if they actually feel threatened by Google's offering. 50,000 - 500,000 people is just not going to be enough to impact any of the big player's bottom lines. Comcast has 15 million Internet customers. Are they really going to mis 50,000 of them?

Here are the Sub numbers for digital tv providers. http://www.ncta.com/Stats/TopMSOs.aspx

That's a good point. When I left they didn't even ask why.

Cool idea, but was anyone else reminded of the Internet2 hype when they read this?

Interesting 2007 Cringley article predicting something like this, except more focused on providing the backbone. http://www.pbs.org/cringely/pulpit/2007/pulpit_20070119_0015... He also claims Google "controls more network fiber than any other organization."

Linking this news to what i read an year ago about google criticizing the router manufacturer for their lousy equipments(performance and both hw/sw quality) i just wonder from who they will buy their optical equipments... Sadly they are not an hw company, so no google router anytime soon.

Supposedly Google designs their own servers and switches already; routers wouldn't be much of a stretch. Or they could call an existing vendor and get customized equipment. At the 50,000-customer level the economy of scale might work.

There seem to be a large number of comments on the possible effects, motives, and general thoughts about Google's effort, but I'm intrigued with what it's going to take to get them to consider a particular community or city. Will they be partial to urban or rural neighborhoods? Densely- or openly-developed communities? Newer or older cities? Will demographics play a role?

I'm sure they're going to be considering all of these, but short of keeping my fingers crossed and filling out an application I hope some part of Austin, TX gets on the map.

Google needs to sell access to entrepreneurs/large entities who want to get into the ISP business. Doing this will create a market over the monopoly/oligopoly we have here in the US.

Reminds me of the Google toilet ISP (http://www.google.com/tisp/).

I don't understand why the go for the 1gbit fiber to the home. it would be more helpfull to their business if everybody in the u.s would have 2 or 5 or 10mbit/sec in a short time , then if everybody has 1gbit/sec in a much longer time.

Also , does anybody know mass market applications that require more then 10mbit up/10mbit down?

Mass market applications that utilise 1gbit connections do not exist because the mass market does not have 1gbit connections.

But connections above 10 mbit/sec exist in many places outside and inside the u.s.

Sure, my connection is advertised as 10 mbit/s.

At 7PM, I'm lucky to get 512kbps.

1 gigabit with relatively low latency would be enough for live jam sessions for serious musicians. I've seen a couple of tele-concerts where the bandwidth between venues was especially set up and I can see some potential there.

Also, multiviewpoint live video feeds, full-size catalog photos for e-commerce (full 12-megapixel photos of the actual product so you can zoom in close enough to see the stitches.).

All this and more.

Comcast is slowly getting into enterprise fiber-based bandwidth for business; my guess is that they will use the enterprise to grow their installed base, then eventually go wholly into fiber based delivery. Of course, at a price... whether this will spur Comcast into a faster rollout is the question.

We'll operate an "open access" network, giving users the choice of multiple service providers.

I'm confused. Won't Google be the sole service provider of its own network?

It sounds like Google will be the "fiber service provider" and you'll have a choice of ISPs.

Forget 5 minutes for a full feature film. If you are getting 1gb/sec., you will be able to download a BlueRay movie in 30-60 sec.

The speeds are 1 gigabit connections. The Dark Knight Blu Ray was approx. 35GiB. At 1 gigabit connections speed it would take 280 seconds in a best case scenario.

That doesn't account for any network overhead and assumes the storage medium can write > 125MiB/s (Pretty much discounting all standard platter drives)

The 5-10 minute figure is reasonable.

Jeez. Back in my day, movies were ~700MB. That way we could put them on these things called CDs. Ever hear of CDs? Damn kids.

Google makes me laugh. They want the world, they want it now, and nobody is going to get in their way. It's great!

Really excited about this. Blazing speed!

interesing, if only google wouldn't bring the cpu speed down to zero by some of its bad javascript code.

Aquinas Router, anyone?...

Yay!!.....hmm..wait a second..

(addendum: whoa, my tail's on fire!!)

I don't understand the down-votes here. This is exactly what I was thinking.

Google Phone, Email, ISP, Website, Checkout, Calendar, Docs, Search, Stats, your DNA .....

People are making a stink about Comcast buying NBC, but Google seems to get away with much more and people cheer. If the argument is "sure, but they make great products", that doesn't cut it for me.

Probably because there's no content to the comment? It doesn't even explain which "hmm" is being hmm'd.

No offense, but there are some comments posted here that may take a second or two of interpretation to understand, perhaps making them that much more witty?

Certainly, but few of them are that vague. Something being witty implies hidden content that further thought will reveal. That one, while perhaps inspiring thought along those lines, contains nothing in and of itself.

Besides. Given the yc crowd and their response to monopolistic habits, I'd be willing to bet that something like 90% or more had the same thought. So you can argue that it's not even inspiring.

"It's a trap!" ~Admiral Ackbar

I for one welcome google as an fiber based internet service provider. Especially if its better than the competitors

I for one welcome google as an fiber based internet service provider. Especially if its better than the competitors I wonder if they will do analytics on the packets or information

This is a very - 1 GIGABIT/SEC - interesting development on part of Google. They've successfully - 1 GIGABIT/SEC - established unified control of - 1 GIGABIT/SEC - their entire chain of service without strangling competition. Where Google goes - 1 GIGABIT/SEC - so does everybody else, it seems - and I, for one, fully welcome our thoroughly broadbanded - 1 GIGABIT/SEC - future.

There seems to be a lot to look forward to. Say, the INCREDIBLE 1 GIGABIT/SEC SPEED AHAHAHAHAAYES

Reply on news.ycombinator.com