Something About Google Fiber Doesn't Sit Right with Me | Hacker News Reader

Something About Google Fiber Doesn't Sit Right with Me

mikecp.net

38 pointsInfinityX044 comments

Ok, if we're going to talk about 'vibes' then lets talk about this one, "Why is someone who is a search engine marketer deconstructing on Google's infrastructure build?" and why is that same article posted to a tech news site by a user with the name of an entrenched infrastructure player? Ok, so it smells like astroturf, but let's set that aside for a moment.

Remember reading all those articles about how SpaceX was 'cheating' or risking safety because their costs were so much less than what 'people who knew how to build rockets' charged? Its a simple rhetoric technique, compare X to Y when X and Y aren't correlated or even have a casual relationship.

Anyway, rather than devolve into some sort of ad-hominem rebuttal, lay out what the costs are for putting fiber into an area. There are easement costs (ok the City covered that), there are facilities (cooling and power for a bunch of switches and routers), IP transit costs and general IP costs (like routable addresses or not). Then there are the sunk costs of installing this stuff.

So one source [1] puts getting fiber to the home at less than $1,000. How many homes are there in KS? The census puts it at about 463,200 [2]. So lets say the average house hold is 2.0 people so that represents 231,700 homes? or $231,700,000 to install fiber to all of them? A lot lower than the estimates put out by the OP.

What is the long term value of customers there? Well we can pretty much assume nobody would be stupid enough to pick Comcast or a bell company for Internet service (10x the cost for 1/100th the bandwidh) so lets say that Internet penetration is 60% [3] so 139 thousand homes, if they got the 'cheap' stuff at $70/month that is $9.73M / month. So how much does Google earn on that revenue? We don't know but we can work it backwards, lets say all up it cost Google 4x what it costs Verizon to lay fiber so $1B, lets further say that fiber is good for 20 years of lifetime. $9.73M$/month times 20 years of 12 months is $2.3B. That is double the investment of $1B back in 20 years. Now not a great rate of return, most would like to see it double twice in 20 years. But not an insane proposition either.

Personally, if ComCast is right and Google is making a huge mistake, I think they should just sit back and watch it happen right? So clearly going to fail at this, their job is done, they can just eat popcorn while the big G sinks under the waves and stops being a thorn in their side. Me though, I would not bet against Google on this one.

[1] http://connectedplanetonline.com/mag/telecom_riding_fttp_cos...

[2] http://www.google.com/publicdata/explore?ds=kf7tgg1uo9ude_&#...

[3] http://arstechnica.com/tech-policy/2009/06/us-20th-in-broadb...

These short-sighted comments really annoy me. I would give parts of my reproductive system to get a fraction of a Google branded fibre and people are getting angsty about what Google gets "in return".

No the sums don't add up. Tax payers will subsidise some of the costs and Google will undoubtedly analyse the traffic to deliver yet even more personalised advertising. The local government will subsidise parts of the installation and make it legally as smooth as possible. Google will probably get cheaper offices or easier planning application processes, and who knows, they'll probably have some sway when it comes to local lawmaking, like any major contributing lobbyist...

... But in return Kansas City gets a bloody gigabit network for pittance. Free ADSL for anybody else.

It's practically unimaginable the services you could deliver to people with that sort of bandwidth because for a very long time it's been considered rude to assume anybody has anything faster than a 2Mbps connection. Google Fibre could stream 4k (with h.265). Even being able to assume the user has some upstream bandwidth for once is a complete boon to developers. In the UK the ratio is 10:1 almost everywhere unless you're on a business line (for four times the price) or the new FTTC options that are available in like 2% of the country.

tl;dr: you get a lot in exchange for letting Google in, not least that you will probably become the testing bed for all sorts of snazzy new technology. Stop moaning, pay your taxes and enjoy being able to watch all the porn at once.

>"No the sums don't add up. Tax payers will subsidise some of the costs and Google will undoubtedly analyse the traffic to deliver yet even more personalised advertising. The local government will subsidise parts of the installation and make it legally as smooth as possible."

As much as I agree with you, this quote highlights something that cycles through my mind whenever I read about Google's ambitions. Basically, Google is doing and motivating people to do what our government should (and as we read, apparently somewhat can in some ways) do to improve our communications infrastructure. It annoys me to think that we could simply be building a new (potentially more neutral) fiber "highway" across the country with a similar model, run by the government. If only people weren't so afraid of government and hell bent on handing things off to private sector instead of the public sector. It's disappointing, and I feel in the very long run it could very well backfire on all of us if google succeeds and becomes the next AT&T.

"It annoys me to think that we could simply be building a new (potentially more neutral) fiber 'highway' across the country with a similar model, run by the government. If only people weren't so afraid of government..."

The government has given us many reasons why we should fear a government-operated internet infrastructure. Do you really want the same government that's seizing web domains without due process to control the physical infrastructure as well? The same government that wants to vacuum up all the internet traffic of law-abiding citizens? That does warrantless wiretaps? That runs the TSA? Be careful what you wish for.

This is incredibly disingenuous I don't know whether to laugh or fucking cry.

We are talking about Kansas City here! A city that has long since been a shadow of its former self -- once a major US hub facilitating cross-statewide trade -- now a landscape of empty derelict factories and office buildings.

At least it was before the City Planners (read: OMG! The Government) did something about it.

For the last few years there has been a huge concerted effect to revive Kansas City and save it from a fate of "Detroitism". Is that so bad? No, it isn't, unless you're of the belief that all of America's cities' outside of LA/NYC should just "suck it up and deal with it" and be the subject of the next "Ruin Porn" hipster tumblr blog.

We are now starting to see some the fruits of these efforts: Sprint Center (Stadium), Kaufmann Center for the Performing Arts and many many other revitalization projects, including Google Fiber.

Spare me.

This isn't about bandwidth. It's about content. The only thing standing between YouTube and the CableTV providers is bandwidth.

If Google cracks the bandwidth nut, they can essentially replace TimeWarner, Comcast, etc. in one fell swoop because they can easily replace 100% of the services those companies offer at a fraction of the cost.

Kansas City is essentially giving Google free reign to perfect their rollout and implementation processes. The next city that Google wires up will probably cost them a fraction of what Kansas City does because I'm certain that Google will automate every last detail that can be automated.

I've got to agree with you an many others in the comments. The thing that "doesn't sit right" is that Google is looking at alternative ways to benefit their company while benefiting the community. There's always going to be a risk of monopoly. In Australia, the fiber roll-out is government owned and funded, replacing a former government telco monopoly with a current government telco monopoly. If the price is right and there's still competition, I think we should be vigilant but happy.

I agree, and I think Google could have more in mind then just replacing television providers (no small feat). They can start to eliminate television channels. Why must NBC, CBS, or ESPN exist? If Google did "adwords" for television they just allow producers to release a show and then Google injects per-user targeted ads. Google's choice of a tablet as a remote hints at more interactive TV, not just big 30-second commercials, but opportunities to buy, learn, tag, etc. while you're watching a show. It gives producers more control over their content and an opportunity to look at different ways to fund the show's production. Forget 30 minutes show slots... do 48, or 10, or whatever makes sense.

Sure, it probably means more ads, but it could mean more innovative, targeted, and available content too.

"To get that in writing, that’s 91.58 Mbps down and 35.01 Mbps up."

No. That's how fast speedtest.net is. Google Fiber is much, much faster. Try it again with http://speedtest.googlefiber.net/

Yeah the 48ms ping was a dead giveaway that something in that test wasn't right. Bad wireless, dodgy connection, poorly chosen server.

An Ethernet→router→fibre→local-speedtest node trip should be closer to 5ms, if that.

Of course Google is "doing this to upset the status quo". The status quo makes a lot of money providing a crappy product with crappy service. Google wants to disrupt that by providing a good product (probably with crappy, but better, service, if they hold true to form) so they can become the new status quo.

And if KC invests in this and can see some knowledge worker jobs come to the Heartland, that seems like a true investment and not a dole.

It's not the "way to good to be true vibe", it's just that people in the US think that they have the best of everything so how can we have something out there so much better for the same. When in reality the US is far behind in infrastructure in a whole. So why this upgrade looks amazing the truth is its the only thing on par with the rest of the developed world.

Define "rest of the world", in most of South America, people get 100kb/s down and 20kb/s up. I guess most of Africa and Asia bar Japan, South Korea and a few others aren't fast either.

I don't know if grandparent post has been edited but it says "rest of the developed world" now.

Lachman's numbers are wrong. Not that those numbers are innaccurate, but that they aren't comparing the right numbers. According to Lachman, the price to connect a single home via fiber is $1250. The price that Google is selling it's service for is $70 a month. If there is no maintenance cost, the return on investment starts happening at month 18 of service. Assuming maintenance costs are a little less than half the installation costs, that gives us around $600 per year.

Cost to Connect: $1250 - $300 = $950

Cost to Maintain/Year: $600

Price/Year: $840

So, assuming that the income chart looks sort of like this, then the point at which it becomes profitable is 4 years. And this assumes that Google doesn't get more efficient.

Google Fiber is supposedly about 10 times faster than that. Presumably, the bandwidth test server was being maxed out. His comparison based on the 91 Mbps result makes current ISPs look better than they should.

> For comparison, where with Comcast I can download a 1080p Breaking Bad episode (1.86 GB) in about a half hour, Google Fiber would get it done in under 3 minutes.

On Google Fiber, that Breaking Bad episode would be downloaded in about 15 seconds at max throughput.

Is it possible the WiFi, or a crappy laptop's hdd, are the culprits for the bottleneck?

I've definitely had speedtest.net results constrained by wifi and a slow laptop that couldn't run the flash control that well.

I get synchronous 200mbps from Webpass for $450 a year. They don't provide content or services on top of that, which means they don't have incentives to analyze my data, nor will they attain perverse infrastructural advantages over other online services. It is the proverbial "dumb pipe" many here wish they had. And I'm very happy with that.

I don't want a world where Google runs my internet connection any more than I want a world where Microsoft, Apple, Cisco, Amazon or Facebook does.

None of this should come as a surprise. Remember when Google was asking cities across America to apply for the privilege of being the first city to have Google Fiber? Cities were competing against each other for Google Fiber.

Google didn't choose Kansas City for no reason. They chose them because KC was willing to be abnormally accommodating in letting Google roll out. After all, if you're going to launch an ambitious, experimental project, you want regulatory red-tape to be the least of your worries.

Wait, I think I know what it is..... Hmm, what's that word.....? Oh yeah.....

Privatization!

Of course it's less than ideal! In a perfect world, we'd have had a nationwide gigabit deployment years ago. Internet would be a global human right, yada yada.

It never occurred to me that profitability was a goal for google fiber. From the start I took it to be a pilot program to understand how people would be using the internet a few years in the future.

I remember an article, way back, which claimed that YouTube was running at a loss, based on a dramatic over-estimation of its costs. This looks like the same thing again: unsupported assertion that Google's costs are going to be huge (because other telcos, which are very wasteful in ways that Google isn't, have huge costs).

I also think this article may be understating the upside, from Google's perspective. It looks like Google wants to displace cable companies not just as ISPs, but also as television distributors, replacing them with YouTube. That'd be worth a lot of money, and it's something they could reasonably achieve - but they need something like Google Fiber to put them in a good bargaining position with the content producers.

I fail to see how any of the things you mentioned are negative.

Taxpayers paying for a commercial service whether they use it or not sounds like one. Special privileges and fast-tracking is another.

You think "taxpayers paying for a commercial service whether they use it or not" sounds like a problem. Other people think this sounds like economic development.

In principle, the first one of those is not intrinsically bad. There are plenty of peripheral benefits to cheap high speed internet that people will reap even if they aren't using it. That's the whole idea behind subsidies.

The second part is what troubles me.

Isn't it very normal for cities to support companies that they want to come into their city?

Whether it's a factory or some other sort of service I don't see what the problem is. If the city government thinks it is a company worth doing what they can to help out that's fine, it's up to their judgement. And if the citizens that live there disagree with the city's government then they should elect new city government when elections come around.

Who in there right mind ever thought there was not a 'make money' angle here. Still want it, still trying to get wife to move to KC.

I'm fairly sure that Google is paying a lot less than $1250 per customer.

1) it costs a lot less to wire a whole neighbourhood than to wire a single house

2) kansas city was chosen because they could run the fiber on telephone poles rather than digging trenches (and they're not paying anything for access to those poles)

Google has stated that the $300 cost they're charging is the cost of wiring up a home, and I believe thats what they think it will cost. There probably will be overages and complications they don't have in their model, but I think that $300 is fairly close.

>Well, then. It seems many Kansas City taxpayers will be paying for Google Fiber whether they sign up or not.

I guess that's why google fibre includes a free service tier.

re:

"It seems many Kansas City taxpayers will be paying for Google Fiber whether they sign up or not. It also seems that Google has a LOT of free and unregulated access in the city. Kansas City could become a testbed for many a Google project."

[because] 'Google received stunning regulatory concessions and incentives.' [quoted within article.]

Yeah. Google got regulatory concessions not available to the 99%. Is that a surprise? This is how "business" is done with local governments across the country. Big businesses routinely get special deals in return for major investment.

Many times, these "investments" don't pan out. Local governments frequently fail to include clawback provisions for those occasions when Scam, Inc., takes the money and runs. It's a bet every time.

As to whether Kansas City taxpayers will pay: the fact is, they have already paid for the infrastructure. (Or they're still paying, via bonds.) So what? Google getting access to rights-of-way sounds like a bet that's better than average.

Proof is in the numbers. (Nice download speeds and the end user costs.) I hope the City Council of New Orleans, LA can get as good a deal.

What are the prices of competitors in the US? Here in New-Brunswick Canada, prices are:

80Mbps down & 30Mbps up: $69.95 /mo. for first 3 months then $100.95 /mo.

50Mbps down & 30Mbps up: $29.95 /mo. for first 3 months then $83.95 /mo.

20Mbps down & 15Mbps up: $29.95 /mo. for first 3 months then $68.95 /mo.

Seems to be not that far from Google's prices.

Not far from their prices - miles from their product. You're comparing 110Mbps for $100/month against 1000Mbps* for $70/month and saying they're near equal.

That doesn't make sense to me.

*I'm judging from a speed test posted on the blog post that the line is asymmetric and down and upstream add up to 1Gbps.

The speedtest result posted in the blog post is really bad. 91.58 Mbps down and 35.01 Mbps up add up to "127 Mbps" in terms of your comparison. Of course, Google Fiber is much faster than that.

For the record, I'm currently paying $30/month (normal price $62/month) for 25/2 DSL in Vancouver, BC. If you live in a wired-up building, you can get 25/10 for $38/month, 50/10 for $62, 100/10 for $83, and 300/15 for $113 from a boutique ISP.

I know Google mentions a "gigabit internet" but in the article they measured "91.58 Mbps down and 35.01 Mbps up".

Comcast, near where I live (in MB/s, all prices are in addtion to a $7/mo modem rental )

3/.768 - $40/mo

20/4 - $30/mo 6 month promotional (incl. cable box rental), $42-62 afterwards (varies)

30/6 - $73/mo

50/10 - $119/mo

105/20 - $199/mo

Comcast seriously needs competition, and I unfortunately don't live in an sonic.net area.

I can't wait for Google to disrupt the hell out of this situation.

Comcast advertises their speeds in Mb/s

Except Google is 1000Mbps down and 1000Mbps up :)

"To get that in writing, that’s 91.58 Mbps down and 35.01 Mbps up. For basically the same price, I’m lucky if I can get 9 Mbps down from Comcast here in Seattle. And that’s probably the fastest residential I’ve had in my life."

I use Comcast in Seattle. I pay $109 a month (for Business Class). Right now, I get 55mbps down, 35mbps up.

And think of the data they can mine by having full access to your Internet usage.

Please read the privacy policy before spreading misinformation:

https://fiber.google.com/legal/privacy.html

Please don't pretend that privacy policies are carved in stone:

http://arstechnica.com/tech-policy/2012/03/googles-new-priva...

It would seem odd if there were multiple utility companies offering to provide the service and Google got an unfair advantage. If nobody else was offering anything then what is wrong with a government agency working with private company to provide service that is helping people?

Reply on news.ycombinator.com