Why Is Google Fiber the Country’s Only Super-Speed Internet? | Hacker News Reader

Why Is Google Fiber the Country’s Only Super-Speed Internet?

wired.com

30 pointssk2code57 comments

Shaming won't work when so many places in the USA only have one viable cable provider. Here in my area of San Diego I'm stuck with Time Warner. Even ignoring the relatively slow speeds, I'd be happy just to not have my connection randomly disappear for about 5 minutes 2 or 3 times a day.

I'd switch to a competitor in a second except there aren't any.

The ISPs have no shame, unless you're willing to force real competition with them nothing will change for those in areas with defacto monopoly ISPs.

I'm assuming that the various ISPs have agreements between each other to not expand onto each others "turf"? Otherwise how does competition not naturally spread? If they do, how do they get away with that?

More puzzling to me is how ISPs squash towns plans to implement their own networks.

I'm very ignorant on how things got so bad over in the US, I've always had a multitude of ISPs over here in England.

I was under the impression that local governments offer them franchise agreements and keep competitors out.

They don't have agreements (which would be illegal), but they don't need them. See: http://en.wikipedia.org/wiki/Natural_monopoly.

When Comcast owns all the fiber in, say Chicago, it's pretty much impossible to encroach on their turf in a cost-effective way. Say Verizon wanted to come in. They'd have to build their own network, and so could never compete with Comcast on price because of the huge capital expenditure required.

The US should do what it does for gas mileage (CAFE Rules). A national roadmap for bandwidth speed (and latency).

This is actually not a bad idea.

Or something similar but maybe even better: Set up a cross-ISP subsidy program. Whoever has the fastest connections in the country receives a subsidy and whoever has the slowest connections has to pay it. Don't even put bandwidth numbers in -- just rank them all by the speed of the unmetered connections they offer for less than $80/month, then the slowest 33% pays a substantial tax to subsidize the fastest 33%. Presto, instant competition to be the fastest, even though they're all in different geographic areas.

That makes absolutely no sense. What would happen is that urban ISP's would win immediately, and then rural ISP's would be stuck basically cementing that lead in place by subsidizing the leaders' capital expenditures.

You're first assuming that there is such a thing as a rural or urban ISP. Have you seen the coverage map for AT&T or Comcast? They're all over the map.

But more than that, there is no reason to think that building a high speed network in a rural area would be impossible rather than merely expensive. So all they have to do is to do it first and they'll be the ones receiving the subsidy. And actually digging the hole in the ground is what costs a lot in rural areas, the terminating equipment doesn't cost anything more just because the wires are longer. So they install fiber to the home once and having paid that cost, the future upgrade costs become equalized with urban areas because all future upgrades come from upgrading the terminating equipment rather than having to put any new wires in the ground.

And if we ever want to have fiber in rural areas then that has to happen eventually anyway, so why put it off?

And, even if you're right, the result is still for urban ISPs to compete with each other to be the fastest and for rural ISPs to compete with each other to not be the slowest. The losers of that competition then have less profit for their shareholders -- or, if they lose very badly, their network is auctioned off to someone with more capital who can make the necessary upgrades. I am not seeing how we can lose here unless "we" are greedy telco shareholders who don't want to pay for upgrades.

One fiber run to a high rise in Chicago hits as many customers as wiring up an entire town in rural Illinois. The idea that ISP's that serve less populated areas could somehow beat ISP's that serve mostly urban areas is completely ridiculous. What would happen is that the urban ISP's would win the race, and the ISP's with low density service areas would just go out of business perpetually paying the subsidy.

It's like trying to incentivize competition in baseball by requiring losing teams to subsidize winning ones. Basically everyone would just be paying to make the Yankee's roster even more iron clad than it is now.

You're just making the argument that the amount of the cross-subsidy would have to be calibrated so that the losers are not bankrupted by it. We care about fairness in baseball because it's a game. Why do we care so much about it if this succeeds in getting the highrises in Chicago gigabit fiber, when up to this point we haven't been able to?

And let's go with it say you're right. Why can't we just create two or three classes of ISPs (like "cars" and "light trucks" for CAFE) based on their coverage density, so that rural and urban ISPs only compete with other rural and urban ISPs respectively?

It's not a matter of fairness it's basic financial calculus. Internet speed is completely a function of capital expenditure. If you take the slower ISP's and siphon away money from them to give to faster ISP's, they'll never have any hope of catching up. At that point, the leading ISP's can just sit om their duff, because the other ISP's are basically paying to cement them in the lead.

This is true even if you separate rural and urban ISP's. Verizon has a great network in the DC area. If they win, then say a Houston ISP subsidizes them. At that point they have zero incentive to improve. Moreover, people in Houston are stuck with even worse service, because money that might be spent on capital expenditures in Houston are going to pad Verizon's profits in DC.

You're overly focused on the incentive part of the equation and ignoring the underlying market dynamics.

>If you take the slower ISP's and siphon away money from them to give to faster ISP's, they'll never have any hope of catching up.

You're assuming that the effect would be that strong that fast -- that it would be large enough that they could never catch up. There is no reason the amount has to be that large.

And the Houston ISP doesn't have to catch Verizon in DC while Verizon is receiving the subsidy, it only has to catch the next-best ISP in the middle third who doesn't get anything and then it can get out of the bottom third who has to pay. Moreover, Verizon still has the incentive to improve even if they're the leader of everyone because the ISP currently at the top of the middle third is angling to get into the top third, and the ISP at the bottom of the top third is trying to get ahead of Verizon to keep themselves from being pushed out when that happens.

I mean what's your alternative? If you want ISPs to pay for upgrades either you reward them for doing better than average or penalize them for doing worse than average or both, and doing both has the added benefit of being revenue neutral. Is there some better alternative, or do you think the entire goal of upgrading connection speeds is unattainable?

I don't think faster Internet is really a pressing national issue, and I think your idea is bizarre and smacks of the same kind of "clever" ideas that always end up disastrous in practice.

If high speed Internet is such a national priority, and the market won't provide it because of natural monopolies, then simply break up the utilities and prevent consolidation.

How do you break up a natural monopoly? They tried it with AT&T along geographical lines and we got what we have now. The expectation was that the Baby Bells would expand into each others' territory and compete with each other instead of buying each other, but if you prohibited them from buying each other, it still doesn't give them any reason to compete with each other instead of just continuing to milk existing infrastructure indefinitely without upgrading and silently agreeing not to expand into anyone's territory who doesn't expand into yours.

In the UK our situation is much better, although BT owns most of the physical cable "Local loop unbundling" forces them to share their lines with competitors.

It seems in the US the only people you can buy from own the physical infrastructure. Once one company has already built their network in an area they form a natural monopoly - it's not economical for competitors to come in and rebuild the network when they know they'll only be able to get a certain % of households to switch.

That used to not be the case with DSL, but about a decade ago the FCC reclassified internet services in such a way that it was no longer the case. The theory was that this would encourage the providers to upgrade the local loop.

[edit] The above was from memory, and it is actually a lot more complicated than that. The decoupling of broadband from voice was challenged in court and the FCC had to change their rules somewhat. I'm not sure what current requirements are. See also:

http://openjurist.org/359/f3d/554/united-states-telecom-asso...

A quick key for reading that: ILEC == people who own the copper CLEC == people who want to use the copper.

Canada and Australia have it worse (bandwidth limits, etc).

Anyhow, the US government should make this a national priority. The faster the internet, the better the economic growth. How can the US continue to dominate (and/or influence) the internet if they do not have the proper infrastructure for it?

A clear message needs to be sent: the telecom/cable duopoly is no longer assured. A "put up" or shut up" attitude needs to be put in play. If they don't like it, then don't take subsidies.

There definitely needs to be more players in order for the market to function properly. I welcome Google or any other to change the game.

I'm living in a small beach town in Vietnam and I can get fiber directly to my house now. That this is easier here than it was for me in Oakland should really be a source of shame.

Dude, fiber in Vietnam is a huge misnomer. You think you're going to get > 20mbps on that new fangled fiber? Good luck getting top shelf DSL speeds.

I had fiber in HCMC. It was a joke.

What keeps the US competitive internet wise is the fact that the commercial internet is separate from residential internet. Google, Apple, Amazon all have among the best $/gbps in the world at their american offices. Having said that, I do agree that faster residential internet would create new markets because of unforeseen possibilities. Just think about it, dropbox would become as fast as a harddrive.

I'm in San Diego too. Do yourself a favor and switch to ATT Uverse Internet. ATT (maximum) gives you a consistent 3MBps versus Time Warners (maximum) 6MBps that averages closer to 1MBps.

Not available in my area. I've looked into the options extensively and the only thing in my exact area other than TWC is AT&T DSL and I'm pretty far from their closest central office here so the speeds would be the absolute minimum.

Everyone here is saying the word monopoly, but realistically ISPs are natural monopolies (http://en.wikipedia.org/wiki/Natural_monopoly). We don't see competition because startups (even incredibly well funded ones) just don't have the funds to start pulling fiber all across cities. This is also why most high-speed internet today is based on some new utilization of older technology: DSL utilizes phone lines and cable utilizes, well, cable — it's even too expensive (and in terms of business, not worth it given their natural monopoly status) for existing companies. This situation will persist for a while, until either Google starts pulling fiber through tunnels in major cities and the threat forces other ISPs to improve. The only other alternative is fast wireless, as then infrastructure investments decrease drastically.

The ISPs that remain are not natural monopolies. My town is exclusively Comcast, but that doesn't mean Comcast beat out the other cable-tv companies on price/service and ran the competition out of town.

Comcast was granted a franchise by the municipality. That's why Comcast spent the money to pull cable all through the town over poles and under trenches. You're granted the monopoly by virtue of spending on the physical plant and providing equal coverage to all sections of town, not just the wealthier neighborhoods that can afford coax-pulling crews.

Google can "shame" the other companies all they want with speed and service but, at least in my town, Comcast doesn't have to do jack shit.

What would you do with more bandwidth?[1]

I've discussed this with several of my friends. In our area, we have multiple providers and can get up to 100MB down/ 30 up with residential service. (You can get all you want if you pay business prices, I work at a company that sells that).

30/5 is the standard. So what would we do with more bandwidth?

Netflix streams in high def. Games from Steam download quickly. My Voip phone works well.

In the future, I'm assuming we will find a use for more bandwidth, but right now, I cannot justify paying for a higher bandwidth connection. (I wouldn't complain if they upgraded our speed for the same price.)

What am I missing? What _could_ I be doing with more bandwidth (besides pirating the entire output of Hollywood and the music and gaming industry)?

By extension then, what argument do we use to force/persuade ISPs and the government (who pays for a lot of infrastructure [2], especially in rural areas) to increase speeds to end users, while maintaining an affordable price structure?

I'm looking for honest answers, I'm not trolling, I promise :-)

[1] Assuming that you already have a relatively fast, stable connection now (somewhere in the 30mb download and 5mb upload range).

[2] BTOP http://www2.ntia.doc.gov/ My father lives on a farm 18 miles from the nearest town. He has fiber.

When I go to visit my family I would rather not have to wait hours to download a piece of software because my sister is watching a hq movie on netflix, my brother is voice chatting with a friend, and my mom is browsing cooking recipes on websites that load a video advertisement every 20 minutes.

With a fast connection could do away with many harddrives. Could pay to purchase cloud storage that could be mounted and used. Wouldnt have to worry about personal backups etc. Would probably still keep a bit of space local but not nearly as much as now.

If you're just asking for point examples, Dropcam (my company) could stream at higher quality (1080p+) and/or support many more cameras per household.

At 1Gbps, cloud storage/backup becomes fast like local storage, but has the benefit of being geographically redundant.

Netflix, for one, can lessen their compression which is actually pretty apparent if you pay attention. You see artifacts everywhere and it looks nowhere as good as the HD torrents that I could get.

Also, transferring files between servers would be less painful. Syncing between servers always takes me a minute or two at least and completely throws off my coding flow. Instead of choosing to host it locally and fix things on the server, I could just quickly sync files and fix things on there.

There are a whole host of reasons:

1) Verizon isn't peddling a lucrative side business to multiply their return on network capital expenditures. Google Fiber is not only making money on the connection fees, but also making money when you use your newly-found bandwidth to put your documents on Google Drive where they mine it and sell your information.

2) The major network providers in the U.S. are the legacy of the telecom utility monopolies. They have enormous networks that are the result of mandates to serve not just Kansas City, but suburban and exurban and rural Kansas. Google has the advantage of starting fresh.

3) ISP's don't have a lot of incentive to do capital investments because they are natural monopolies (in addition to being the legacy of utility monopolies). So you need government involvement to introduce competition in one way or another. However, there are no votes for such involvement. As a practical matter, super speed internet is something that only makes economic sense right now to build out in urban areas. But districts in the U.S. are heavily gerry-mandered to the point where rural votes count for much more than urban votes. At the governmental level, if you want to bring super speed internet to Kansas City, the whole debate will get bogged down in "but what about rural Kansas?"

We look enviously at South Korea's high speed internet, but nobody ever points out that more than half of the country lives in Seoul If everyone in the U.S. lived in NYC we'd probably have awesome internet here.

FiOS is already at the point where the bottleneck is usually the wifi network not the internet connection. The market for "tethered to your desk" home internet is probably not that big.

I think the problem is rather that "FiOS" is not the class of internet connection that most places in the United States currently have.

My point is that "super speed" Internet is pretty useless when wifi networks max out at 20-40 Mbps in practice. And while not everyone has access to 30 Mbps+ Internet, most people in urban areas do.

Wifi networks max out at those speeds because "super speed" internet is elusive and so there is insufficient demand for faster/better wireless devices. 802.11n supports significantly higher speeds than that already, but nobody uses it because 802.11g is faster than most internet connections. Moreover, we're talking about building networks that haven't been built yet. By the time we build them there will likely be 802.11ac/802.11ad devices on the market that can handle even more.

802.11(n), with three streams and a wide channel just barely covers my 75/35 FiOS connection, in one direction at a time only.

Take a look at the speed tests on, for example, SmallNetBuilder. "450mbps" 802.11n gear does less than 100mbps—and that's half-duplex.

11(ac) gear will do better, at the cost of using up a ridiculous amount of the 5GHz spectrum.

The speed listed on the box is, well, what most of would call a lie.

In the US it's often really hard to establish any kind of cabling network due to right of way laws. This appears to be the worse in metro areas with the richest NIMBYs such as San Francisco or Palo Alto despite having the large populations interested in super-speed internet.

I have Comcast Business 50/10 at home -- it's pretty decent, but there is some crazy Comcast routing going on, so sometimes services like Netflix fail badly. With a VPN to a local server, it's pretty decent, especially since I pay around $60/mo due to a legacy discount (otherwise $200+/mo)

What I really want is fast, reliable, dumb-pipe bandwidth from various locations to my datacenter, so I can do things like back up at ~100MB/sec to a SAN in the colo, move VMs around, etc. The only way that seems likely to happen is if I live immediately adjacent and run fiber, or find someplace in the downtown core of Palo Alto with metro fiber and get on there.

It's not actually, but ...

I work at Penn State (one of the Internet2 participants) and have blazing fast internet at work. I've clocked it as high as 485MBps. When I go home (to a small village 4 miles from the campus), I have to suffer with Verizon DSL, Comcast cable modem service or a similarly rated line of sight radio service.

Verizon planned to install FiOS in central PA, but the FiOS roll-out has been cancelled. Why can't I have high-speed and affordable Internet at home? Because the ISPs don't want me to get used to it. Nor do they want me to know how much they've been overcharging me the last 15 years.

At least the US isn't as bad as Canada where you can pay $225 a month for 250mbit down and up.

http://www.rogers.com/web/link/hispeedBrowseFlowDefaultPlans

edit: a bit of searching shows that FIOS has a similar price for their top plan. On the other hand, we have a 500GB data cap. Smaller plans have caps that are lower than 100GB.

I have 100Mb/100Mb for $45/month which is pretty decent I think. It's not 1Gb, but still beats the pants off everything else in the area.

what provider and area if you don't mind me asking?

I'm stuck with around 15/10 for $60 sadly.

I am stuck with 5 Mbps for $50 sadly.

That is really good. What location and what provider, if you don't mind me asking?

UTOPIA in Utah with the ISP being XMission

I have 40 mbit through CenturyLink (DSL) in Utah as well. Not as awesome as fiber (inconsistent latency), but the competition from UTOPIA/iProvo does seem to drive speeds higher. Wish this were more common in other parts of the country.

I only pay 30/mo (35 or so after modem rental) on a promotion, 45 or so afterwords without contract. I'm thinking of switching to veracity unless I move to a UTOPIA city.

Good times in Utah...

Veracity is a UTOPIA provider as well. I had them for a bit and was happy with them, but switched back to XMission due to price, but now they have the same price as XMission. It's worth pointing out that I switched from XMission to Veracity because Veracity had a much better price to start with.

I think maybe the ISPs don't care. My u-verse modem is running something like 5mbps up, but they won't sell me more than 1.5. They have the infrastructure in place, at an existing customer and they cant be bothered to take my money.

I just had to install a cable line to get 4mbps up for a project.

http://www.webpass.net/

http://www.engadget.com/2011/06/13/sonic-net-starts-trial-of...

Have they actually offered 1 Gbps to anyone yet? Last I read about them about a year ago they were "only" offering up to 300 Mbps, I think.

http://arstechnica.com/tech-policy/2012/02/gigabit-internet-...

All those veritable mountains of wireless routers spread across every city but can we build an easy to use open source mesh network? <samuelljackson> Hell no! </samuelljackson>

From the article: >Computers & Tele-Comm, Inc. (CTC) has been offering a 1 gigabit wireless service in Kansas City since before Google Fiber

Why Kansas City is getting all the love?

Monopolies.

tl;dr: Money.

It's not. Burlington Vermont's City owned "Burlington Telecom" offers gigabit symmetrical.

http://www.burlingtontelecom.net/Residential/Internet/Broadb...

tl; dr: Google wins.

Reply on news.ycombinator.com