Hacker News new | past | comments | ask | show | jobs | submit login
Project Loon (google.com)
1094 points by sabalaba on June 15, 2013 | hide | past | favorite | 286 comments

There's a lot that I don't like about Google, but I do like that they're fucking insane.

This perspective made me laugh and earned my honest upvote; especially because it sheds a different light on Google Glass (about which I have pretty negative feelings).

Well said :)

Is it really a good idea to use Google even on the lowest layer, meaning the network itself, for the transportation of your data?

In light of the surveillance state that the US has become and the collaboration of government agencies with the tech giants, I say:

Support the small guys!

What makes you think the _small_ guys will protect your data _better_?

Because they are minor targets.

Google seems to be the only company that has even tried to stand up against surveillance.

Wasn't that yahoo?

"I love those who yearn for the impossible." -Goethe

I wish other large companies showed such ambition. Telecoms like AT&T can't even be bothered to roll out the network upgrades they promised in the early 2000's.

Most large companies don't have the financial flexibility that Google's near-monopoly on search offers. Lots of ten-billion per year in revenue companies are eking out 5% margins in extremely competitive fields and can't afford to be distracted by ambitious projects. AT&T is locked in tight competition with Verizon in a very capital-intensive market, and can't afford to divert capital to blue-sky projects.

Back when AT&T was a monopoly, we did see innovative projects like this come out of them. At the turn of the century, linking everyone in the U.S. with phone lines was a pretty blue-sky project. The AT&T phone network was a marvel of engineering of the time. And of course Bell Labs gave us tons of innovations.

That's the beautiful thing about being in a market where you're not scraping around in the dirt trying to claw out the eyes of your competitors.

That's pretty weak defense of AT&T. They have 250,000+ employees. They spent half a billion dollors in alternative fuel vehicles investments in 2011 alone. They have 20B of total capital investment a year. Blue sky projects such as Loon typically takes couple of dozen employees and may be few million dollars to kick off - small enough to be rounding errors in AT&T's budgets. If I have to imagine, the reason AT&T doesn't do this is probably because (1) their leadership has lost habits of innovative/creative envisioning (2) their leadership probably laughed this off because potential for revenue from people in Sahara was not very tempting (3) Their middle management is not empowered or willing enough to paddle such ideas to top in the first place (4) They simply doesn't have setup where bottom-up vision is possible anymore so unless leadership envisions it and executes, it doesn't get done.

The lack of money and employees is probably the last thing that would prevent AT&T to do these kind of projects.

This financial flexibility certainly helps, but Google is hardly the only tech company with loads of cash. Consider Apple. They have famously high margins and insane cash reserves, projected to grow beyond $170 billion in 2013[1]. And to caricature a bit, they are dead focused on creating the most comfortable sofa browsing experience for rich people. This focus is good, it yields amazing products, but I do give Google credit for investing in such a crazy project as Loon, with the potential to benefit the poorest regions of the world.

[1] http://paritynews.com/business/item/847-moody-apples-cash-re...

Financial flexibility is nothing.

For most organizations the task of running these sort of experiments will easily come at a chicken feed cost.

However the real problem is really amidst all red tape, most middle management can't tolerate smart engineers. It threatens them, they feel insecure that smart people will ultimately replace them.

There fore they make all attempts to stop anything big from ever happening. Running daily affairs as usual is OK. Its in their interest to keep rewards, motivation and level of work such that their positions are safe.

Minor nitpick: AT&T has done a shit-ton of research: http://en.wikipedia.org/wiki/Bell_Labs

edit: apparently a different beast than today's AT&T

Minor nitpick: The company now known as AT&T is essentially Cingular wireless; they bought the remnants of AT&Ts cell network and the rights to the name.

This explains clearly what happened with AT&T, and its multiple acquisitions over the years:


There's a lot more to it than the wireless piece. Today's AT&T was Southwestern Bell/SBC which as you said, acquired AT&T and switched names.

Thanks, didn't know.

Actually that is the company now known as AT&T Wireless. The company now known as AT&T was one of the 24 'Baby Bells' that were created when antitrust regulators broke up AT&T (the one that created Bell Labs) in 1984.

The company used to be known as SBC Communications but changed it's name to AT&T after acquiring what was left of AT&T (mostly long-distance phone services) in 2005.

Through a long series of acquisitions AT&T now owns 11/24 Baby Bells.



[2]My grandfather worked for AT&T for his entire career and I've involuntarily listened to a minimum of 100 hours of company history throughout my life.

Colbert explains it in detail: http://www.ebaumsworld.com/video/watch/955486/

ebaumsworld is still slapping their watermark on anything that moves?

And the now AT&T carried over SBC's senior management, with all their -- in my opinion -- bad attitude and practices.

Today's AT&T is not the AT&T of yore. Neither in its technical research, nor in its management.

*its name

AT&T spun off Bell Labs in the 90's as part of Lucent. Lucent was acquired by Alcatel. Alcatel shut down Bell labs in 2008.

And of the computer science and other researchers that were at Bell Labs at the time, many now work at Google.

And then they started patent trolling with the trove of patents they acquired.

Doesn't say that on Wikipedia, did I miss something? Looks to me like they're just shifting Bell Labs' focus from basic scientific research and focusing all research into computer science aspects. https://en.wikipedia.org/wiki/Bell_Labs#2000s

AT&T also started and continued "AT&T labs" in 1996, which still exists, and was populated by many ex-Bell Labs folks.

That's not quite the same AT&T that everyone knows of today...

Bell Labs is long gone, and existed to appease the government about a monopoly.

This is about BIG thinking enabled by the unique org structure. Everything about corporate governance says this is a bad structure. But they still might pull this greatness off!

If any technology that was formed in AT&T labs that threaten AT&T's very survival or business model, they will certainly or did squash it.

And little things like Unix

and C

How long does a 100k+ employee company get credit for something a small handful of their employees did almost half a century ago? Not that C wasn't an achievement but . . .

The old monopoly AT&T wasn't just a company but presented itself as an American institution. They heavily publicized their Bell Labs inventions to help keep the anti-trust regulators off their back. Unix/C certainly benefited from this.

I take my point back as the AT&T of today is not the same as it was back then, but the fact that the company funded this research is what makes it admirable.

I'm sorry. What have you done for me lately? C and Unix don't count.

AT&T was planning to connect every home in America with fiber by the year 2000. Then they were broken up and the project was impossible to complete. The current AT&T is nothing like the old one. It's not even fair to compare them

Google deserves a lot fo credit but let's give something to Jerry Knoblach and his team:

"His plan: Beam it down from balloons hovering at the edge of space.

This isn't just hot air. His company, Space Data Corp., already launches 10 balloons a day across the Southern U.S., providing specialized telecom services to truckers and oil companies. His balloons soar 20 miles into the stratosphere, each carrying a shoebox-size payload of electronics that acts like a mini cellphone "tower" covering thousands of square miles below.

His idea has caught the eye of Google Inc., GOOG -0.22% according to people familiar with the matter. The Internet giant -- which is now pushing into wireless services -- has considered contracting with Space Data or even buying the firm, according to one person." http://online.wsj.com/public/article/SB120347353988378955.ht...

Thanks for the history. For those only reading the excerpt, this article is from 2008.


"I love those who yearn for the impossible." -Goethe

Google is taking this global, but I believe they bought the company behind this. They have been working on this for over 10 years!

" But Space Data says its plan to create America's first floating wireless network -- by putting disposable transmitters on government weather balloons -- has already undergone successful testing and is economically viable.

A trial run with text-messaging service in the Phoenix area is slated for this summer. The official launch of the messaging service would begin next spring in Arizona, New Mexico, Texas and Oklahoma, extending nationwide by the end of 2003. Cellular phone service could be added a year later if the company secures more funding. " http://emoglen.law.columbia.edu/LIS/archive/wireless-tech/in...

At&T has no incentive, this probably would help new incumbents and they'll hate that. Google is not going to do something that helps a search competitor, so why should AT&T helps theirs?

It's fascinating to see how well Google has branched out. It started as just a search engine competing with Lycos, AltaVista, & Yahoo and yet they didn't just stop, they kept going. They built all these new services & products around their core product (search) and it's paid off tremendously.

It's definitely something to model future companies off of.

When you win the race, don't stop, keep running.

Building computer networks and artificial intelligence are pretty much Google's core competencies. Most if not all other companies would struggle to fund and develop this project.

First of all, the system is in total very expensive: You don't get any coverage with just a few baloons, you probably need hundreds to make any of them stay connected to the rest of the internet most of the time.

Second, this system is extremely stochastic and complex. The winds are mildly predictable, so they will have to optimize the balloon paths to cover their userbase.

This is a very exciting problem once you think about it...

> This is a very exciting problem once you think about it...

That's a good summary of why Google has been so successful.

They have a way of making engineers and regular people really excited about their projects. It's a feedback loop - have exciting stuff, recruit smarter people, work on more exciting stuff.

Google has many exciting projects, but those are not responsible for its success. Google's success is owed to comparatively pedestrian products: a search engine and an ad brokerage.

Take away Glass, self driving cars, Loon, and all the other blue sky projects, and you still have a wildly successful company. But take away those two cash cows and Google is broke.

I think you are misunderstanding google. For one thing, search engines and their Ad system is the same class of problem as the Loon project: Large Scale stochastic computation.

On the other hand there is no telling what sorts of massive successes will come out of Google. They will have their stakes in any mass-produced self-driving car, Google Glass is at least going somewhere, and Android is already very popular and enabling Google to get Ad revenue without being fought by Apple every inch of their way.

The stock price and market capitalization is driven both by current revenue but more so by these insane projects and the probability that there will be another big success.

Actually it seems cheap to scale. You need a lot of very very smart software but google have a couple of hundred idling geniuses to spare. And if you change helium with hydrogen (not a big deal if a balloon blows up) - everything else is standard components - i think that if they aim for third world coverage the price of transmission must be cheaper than current wireless. And you got much less regulatory hurdles.

Almost everything they're done is centered around ads and getting more data on their users to sell more ads. Even this project, in the end it's about getting more people online to look at their ads. (not sure how much ad revenue you can get from poor places, but whatever)

EDIT: To those saying they're playing the long game; Playing the long game doesn't make sense. The long game is being fully wired. By the time poor areas become economically important enough to be advertised to, they will have traditional telecoms servicing them.

The long game is being fully wired.

No, the long game is wireless communications (and renewable energy). One of the reasons for fast growth in some developing countries is the fact that they don't have to make the same investments in infrastructure that developed countries made in the past.

Wireless is a hack, it will never replace wired. Sure, in Africa they get online with mobile phones, but that's only because it's not profitable _yet_ to run landlines. Even 56K would be better than mobile, but since it's Africa they don't even have that.

Give it a few more decades and they'll be wired just like everyone else.

Particularly over long, sparse networks, wired connectivity is not as cost effective as wireless connectivity, even including the cost of pole assets and antenna mounts.

And data rates of > 1 Gigabit for wireless connections are readily achievable these days. [1]

[1] http://www.bridgewave.com/solutions/whitepapers.cfm

The United States is a long, sparse country, one of the longest, sparsest countries on the globe. We are wired because running backbones over 3G blows.

People say infrastructure is too expensive to build, it's not. When you have a discerning customer base with the purchasing power to demand it, infrastructure is the only way to serve it. Wireless cannot support that. As soon as the developing world stops being developing, in will come the telcos to build real kit.

So, a couple things here.

#1 - The United States is wired for legacy reasons, our infrastructure was built 50+ years ago when wireless voice and data for the most part did not exist.

#2 - When I say "sparse" I'm referring to density of the target population, as in, the number of households per square mile.

It does make sense to run wired connections to dense urban populations, the cost-per-connection, particularly measured on a bandwidth-per-drop-per-dollar basis is great. Also, if you already have cable (either pairs of copper, or coax-for-TV) running to those people's homes, you've got a foundation for delivering data.

For most of the developing world in the next century (excluding the cities), communications, data, and entertainment will be in the form of wireless (Satellite, 3G, or directional 50 GHz+ data drops for backbones).

We won't be running cables to supply communities in the African countryside with high speed Internet. (We also won't be running it over 3G/LTE - which wouldn't be most people's choice for a wireless data backbone these days)

You have no idea what you are talking about. I live in a developing country (bangladesh) which is mostly poor. About 80-90% of internet users are wireless internet users, and every year its getting more dominating.

Wired connection is extremely expensive to deploy and very time consuming too, wireless, not so much. As a matter of fact there is no "nation-wide" wired internet service here, only local providers. New wireless providers are popping up almost every year...

> Wireless is a hack

It's been years since I've had a landline phone. I'm sure I'll eventually drop my landline high speed internet connection too.

It's a hack for the use case of basic connectivity. Your cell phone connects to a cell tower, which in turn connects to a trunk line. That's the proper way to do wireless.

In the developing world, cell towers connect to other cell towers wirelessly. That's a hack. It's the reason why hotel Internet blows. Not its proper use case, engineering-wise. Service sucks but the customer base isn't discerning enough to notice and wouldn't be able to afford better.

We have Clear Internet service in Atlanta. We had it for awhile, but crappy reception due to trees even though we were well covered by three towers made us go back to cable. Dialup would have been better. I had another friend within eyesight of a tower, he dropped it too because service got so crowded and bad that he couldn't stand it.

Current wireless technology could never replace wired and deliver the same quality of service that wires do. For proof, head down to your local Starbucks and try to get work done. I've not seen one yet where the connection didn't cut out at least once an hour. Radios degrade over time, copper doesn't.

Maybe you just had a bad experience with your provider in Atlanta. Webpass in SF is pretty awesome. 200mbps for 50/mo.

Clear Internet is 4G. He was giving an example of how "wireless is a hack". As far as I can tell from their website, Webpass is wired.

How is 56K better than 4G or at least 3G?

>Sure, in Africa they get online with mobile phones, but that's only because it's not profitable _yet_ to run landlines.

I'm not sure it will be ever be profitable to run landlines. The big issue being that there wont never be public and private monopolies again and that you dont run a line to a single costumer.

I feel bad for people like you.

I fee bad that you can't believe that some people or company with money just wants to make the world a better place.

You said it yourself. Google probably won't make money. Maybe, just maybe, they just want to do something good.

Why can't they do both? Make the world a better place and make a nice profit while doing so?

Exactly. That is exactly why I like Google's business model. It allows them to earn money, lots of it, while still being good, at least in theory.

except that anything they do in good faith, could be easily and mechanically turned into more profit by doing something a bit more evil. The pressure to make profit is the driving force, and so the stable norm for any profit making entity is to "turn evil".

being a public company, Google can't really justify altruistic endeavours to its shareholders unless they will also benefit from the project.

With Google's share class structure, they don't have to justify anything to their shareholders, only to their employees.

Whoa?! Care to elaborate on that or at least point to a relevant link? You really made me curious...

He's slightly exaggerating, IIRC, but it's conceptually correct.

Regarding the share structure, Google has two classes of shares, or maybe three by now. Class C shares don't get to vote, and I can't remember if they exist yet (IIRC they were voted into existence a year ago, but only pending various hurdles that I think might not have yet been cleared). Class B shares are worth 10x as much as Class A shares, for voting purposes. Guess who owns most (all?) of the Class B shares.

Anyway, the various share classes are all a bit beside-the-point. The real point is that when you add up all the votes, Larry and Sergey control a lot of them. And then there's Eric. And then there are a few CxOs/SVPs with meaningful shares. So basically a small group of people, who have led Google to its current state, are in control. Other shareholders are just along for the ride.

On top of all that, Google has always warned potential buyers that they don't give a shit about quarterly earnings, and they care about long plays and such, from the IPO forward. If minority shareholders were to sue the company for lack of fiduciary care, I think this would go poorly for them.

You could look up modern SEC filings to have a more-detailed breakdown, but here's a years-out-of-date summary I found with a few minutes use of... you guessed it.


Its the classic News Corp style founder control non-capitalism stock market listing. Other shareholders are not owners, just along for the ride. Illegal in most countries other than the US, where shareholders actually have to own a public company...

Class C shares don't exist yet...they're structured as essentially a stock split, where each class A shareholder gets an equivalent class C share that's non-voting, and so you'll see Google's stock price roughly halve when the transaction goes live.

Between them, Larry Page, Sergey Brin, and Eric Schmidt hold a majority of Google shareholders' voting power (about 2/3 is what I usually hear). I won't link anything because I don't really know the authoritative source for this sort of information, but Googling [Google stock structure] turns up plenty of secondary sources.

without addressing the remainder of your comment..

> I feel bad that you can't believe [..]

the same could be said for naiveté.

Contrast their behavior, with other companies sitting on mountains of cash.

They don't need to do this. Yet they do. That's good enough for me.

Though opening up their index and creating a search marketplace might seal the deal.

Whoa, they'd rather do a thousand as many baloons as whatever amount they're doing now, than that last part of your post. That's, you know, core business.


But think about it...right now they have a random bunch of 10 or so tabs (flights, recipes, patents etc) above the search box. Is that it?

To expand search to its full potential that list has to expand.

What better way to do it than create a marketplace? It would be a good experiment to run a pilot in some small country or city and see what happens. Unlimited API access to google search...any app built on top of that platform that gets X number of hits becomes another tab above your search bar...for a price...or advertisers could make it free cause domain based search can get them better targeting.

There's cynicism on one end and naivety on the other, but there's a whole spread in between. Despite what the cynic thinks, reality is almost always somewhere in the middle there.

It's not like in mid-90's the Internet started getting big and the telecoms started running wires because it was the best option. They based the new network offerings on the infrastructure that had been in place for decades.

Developing areas don't have that existing infrastructure. Perhaps setting up a wired infrastructure will make the most financial sense. However, I don't think that's a foregone conclusion, and a pure wireless infrastructure may be far cheaper and more profitable.

I think the key is that wired networks were ideal in the situation where customers would pay off the investment. Wireless is way cheaper - one station could be 20+ people and take far less investment upfront.

This is not really borne out in the economics. Wireless is only cheaper if you have a big wired trunk line somewhere you can drop a tower on.

It's not cheaper if you have to build that trunk line, and the latency and equipment start getting costly when you're looking at going tower-to-tower over a long distance, since with a single fiber you can run over 100km before you need another relay (and carry a lot more data).

Where this initiative by Google works is getting the rest of the world onto the internet somehow, and profiting off being the base-level service provider. Which would be quite a thing - a world where everyone could send some packet data at any time, would be very different to today where even leaving my house raises serious questions about how online I maybe over the next 48 hours.

>not sure how much ad revenue you can get from poor places, but whatever

They're playing the long game. They see the third world as a market in the making, and they want to be entrenched when it develops.

The goal is to make the world a better place, advertising is just the means to do so.

Even if its true, it's still a great win-win for everybody then.

They're thinking long term. If you provide cheap (free?) balloon-based internet access, you will open opportunities for people who have little access to the outside world.

> The long game is being fully wired. By the time poor > areas become economically important enough to be > advertised to, they will have traditional telecoms > servicing them.

... And by that time, they will all be users of the google ecosystem thanks to this.

Joke: Maybe they have some incentive from NSA or whatever to get and store information from everyone on this planet.

considering the news, makes sense they'd do this

True. They aims to make Chrome OS more popular. It only makes sense if Internet access is available everywhere.

I'd say it still hasn't branched out. It is very much an "internet" company. So anything related to the internet (and data nowadays) is somehow going to benefit their bottom-line.

That said, no internet company has products as diverse as Google's? Smartphone OS, wearable computing, gigabit internet etc.

Autonomous cars, too.

autonomous cars mean people can be on the internet while driving.

I can imagine if Space X was Google it would be stated that the only reason for the project is that they want to serve ads in space.

You can make that connection with anything any company does.

Yes, like a box making company. Or a johnny-on-the-spot one.

I think the phrase "to be on the internet" is going to finally loose its meaning soon. With everything around you (and on you) connected to the global network, you'll be on the Internet all the time. Just as you have your eyes and ears open all the time.

But they also do so much more than that.

It's fascinating to see how well Google has branched out. It started as just a search engine competing with Lycos, AltaVista, & Yahoo and yet they didn't just stop, they kept going.

This project is not a branching out. This is what happens when it's cheaper to expand your market (create the pie bigger) than trying to cutting into market share (take a bigger bite out of the pie).

Driverless cars is no different.

it's not branching out as much as broadening their distribution channels. every project meant to increase availability of internet access is increasing google ad traffic, but hey, this one is crazy and amazing!

And now they sell your data?

No, they sell ads targeted at you based on your data. Nobody but Google and the gov't has access to any of your data on their servers.

That's odd, I'm looking at my calendar and it doesn't say April 1st.

That's the first thing I thought as well!!

My first thought as well :) ... I believed it only because we're nowhere near April :P

It's so encouraging to see a Hacker News comment section that is almost overwhelming informative and positive. I get varying amounts of flak for constantly posting about the quality and nature of HN comments, but as an unabashed optimist it means a lot to me to me.

Yeah, I'm looking for a discussion about technology and maybe somebody have an idea how latency would be. But all I see is typical comments you find in any social network.

Eric Brewer, who joined Google a year or so back, has been working on Technology and Infrastructure for Emerging Regions. He had similar ideas.

I wonder if he has been involved with this project.

http://www.cs.berkeley.edu/~brewer/ http://tier.cs.berkeley.edu/drupal/

Eric is an altruistic guy and a great scientist. I'm sure he's building great stuff at his Google sabbatical!

I believe he is. I've followed his research work for quite a while and Ninja Project seems very similar to Project Loon.

I really enjoyed the subtle design / aspects of the loon landing page.

Starting with the darker background of the thermosphere and progressing into the twilight of the troposphere was a really nice touch. I also just noticed the small altimeter on the left of the page.

Also the whole phonetic wordplay with "loon" and "lunar" was cute. Coming from New Hampshire, I naturally immediately thought of the aquatic bird:


But the "lunar" aspect wasn't lost on me.

Kudos, this type of subtle design is refreshing in the wake of the iOS7 release.

I immediately thought of "loon" as part of "balloon" and also "loon" as in insane or ridiculous. It's a perfect name for something so ridiculous with the potential to be ridiculously awesome.

This is an awesome idea.

However, in my mind, there is a potential issue that the article doesn't seem to mention.

How do they keep the balloons up for a long period of time, fighting all of the different variables of weather, gas leakage, temperature variances, material degradation due to being outdoors?

Fiber and copper stays in the ground undisturbed for decades, because it's cheap & low maintenance, barring from breaks.

While this seems cheap and easy to deploy, keeping it running is another thing entirely

I'm assuming there's some sort of monitoring onboard system and when they need new parts, they'd just land at one of these predetermined stations and be refitted. There weren't a lot of details in that video but it seemed to me like they implied they were inexpensive and easy to swap parts from.

More likely they will not even bother to fix broken balloons and just launch new ones.

Sounds like they're just going to be replacing them all a few times a year: each balloon is expected to "stay aloft for 100+ days at a time."

cite: http://www.google.com/loon/faq/

"Project Loon balloons float in the stratosphere, twice as high as airplanes and the weather."

But they're tethered, right? Or no? if they're untethered, that's just crazy awesome. If they're tethered, I'm still slightly worried about the cables.

OH! They're just going to make enough that there is always one overhead! That's so damn cool. I can barely believe it. Wireless technology is some of the most amazing stuff.

Not tethered -- watch http://www.google.com/loon/how/

Not tethered, as nostromo mentioned. The hazards created by cables in the air for airplanes would kill the idea before it could take off (pun not intended).

Definitely not tethered. Cables become super heavy at those lengths. Not to mention you need very strong/thick cables.

If you just try flying a kite on a good windy day an keep feeding it string, it will pretty quickly start loosing altitude.

Laying fiber to non-dense populations is usually prohibitively expensive.

Wow....this is mind-blowing.

The potential impact of a globally connected internet is crazy. No longer can a country fully filter/control all internet traffic.

This could be a major boon to democracy.

I wonder if Google did this, partially to piss China off.

This seems like an awesome way to get around the 'Great Firewall'.

I wonder how long it'll be before reports surface of authoritarian countries deploying anti-Loon lasers.

In all seriousness, I'm curious how difficult it would be to jam this network.

It's low power in known bands, but coving a very large area. Maybe somebody could launch their own jamming balloons. In the short term Google says they'll abide local laws so I suspect no jamming will be needed.

>I wonder if Google did this, partially to piss China off.

Clearly, not. China has enough air power (including drones) to shut down any amount of the balloons.

I don't think conventional drones can fly that high.

Begun, the balloon wars have.

Actually, kite fights are an old Asian tradition: http://answers.yahoo.com/question/index?qid=20101010110043AA...

Missiles can.

Missiles that can are much more expensive than baloons

If I'm China, I don't care about how expensive the targets are, I just care whether keeping my death-grip on the Internet is worth the price of the ammunition.

Then how about baloons with guns?

>This seems like an awesome way to get around the 'Great Firewall'.

Who says you can deploy them over china?

>Who says you can deploy them over china?

Who says you can't? As far as it seems, it will be circling Earth in low-orbit. Not sure China could stop them if they wanted to.

Previous discussion:


The amazing thing is this has been a rumor[0] tossed around for years (2008).

[0] http://www.techdirt.com/articles/20080220/123009308.shtml

Luckily it's not a rumor any more, Look at the article's url!

I'm not sure what you mean. The previous conversations, and history of stories, offer insite into the technical, cultural, and legal issues that everyone might find valuable. News without context is just hype.

> Loon balloons are superpressure, which enable them to stay aloft for 100+ days at a time. This is far longer than typical weather balloons, which last for a matter of hours. Loon balloons are also unique in that they are steerable and entirely solar powered.

It's like they are putting up cheap, disposable, steerable satellites up in the stratosphere! Genius.

Would be pretty interesting to understand the monitoring and control system for the balloons - not sure if they can bring back the failing balloons to a maintenance site to refurbish and relaunch them or they just dump the debris along with the failed balloon. I suppose at least the electronics aboard must be reusable for a time longer than the 100+ days the balloon stays afloat.

Trying to do an "Energy Budget" with my low-power research background. Please help with the balloon angle.

.1 - 1W Balloon height adjustments

.1 - .5W Flight tracking, control and management

.5 - 2W RF power and conversions (balloon-to-ground)

.5 - 2W RF power and conversions (balloon-to-baloon)

1 - 2W Packet routing, processing and baseband processing

Essential missing insight: how to they lower/rise the balloons and deflate/inflate them? With 100days service time you cannot simply vent gas. So a pump and compressor is needed?

Did it say somewhere that they last 100 days? That number sounds familiar, but I couldn't find it just now when I went back to the site. I wonder how far the balloons would travel during those 100 days. As you can tell from the video, they're going to spend most of their lives drifting over the ocean.

Edit: there are more details here: https://news.ycombinator.com/item?id=5887330 Apparently they circle the Earth every 14 days.

If you just have a radio repeater you could ditch the packet routing.

This is fucking nuts. I feel weird. So this is what real innovation feels like.

This sounded like a great idea to me at first. But,

Each balloon provides coverage to a 40km diameter area. That's a lot of balloons to cover rural Africa. It would help a great deal if the balloons could be made to stay in precise locations, like geostationary satellites. I thought maybe they had solved that problem (which would be amazing). But, I guess not. They said in the video that the balloons would follow the wind all the way around the Earth. Meaning, at any given time, the overwhelming majority of balloons are over the ocean. The difficulty, then, is not merely putting enough balloons up to cover the ground in 40km increments, but covering the ocean as well, in a solid band around the entire Earth. It's not hard to imagine that a stupendous number of balloons would be required to provide any meaningful level of coverage.

I think this is more Google hubris than technical innovation.

The balloons are not going to move in one direction though. The video also mentioned you can steer the balloons, which means, there is way to keep them above the ground and not the ocean.

If you watch the video, they say that the winds mostly blow west-to-east, opposite the Earth's rotation, and they can't counteract that. They show the balloons traveling across the ocean.

It doesn't help the way the landmasses are oriented in the southern hemisphere. I think it's a given that the majority of balloons will always be over water. They mentioned a demo in, I think, New Zealand with a handful of balloons. It would be interesting to know how long they'll have for the demo before the balloons are pushed out to sea, and whether the coverage area will be constantly drifting.

The video seemed to show the balloons in a thin band encircling the Earth. Aiming to cover a thin band would seem far likelier to be achievable than spreading them over a broad swath. They also mention using the steering ability to form them in clusters, which is probably what you'd need to solidly cover an area.

Just cover the continent will require a lot of balloons. The world has 148,429,000 sq km. If you cover 1600 sq km per balloon, that is 92768 balloons. Just cover Africa will require around 20k balloons. It will only work out, if Google can increase the coverage area per balloon.

20K ballons doesn't seem like that many unless each one costs a ton of money. From an industrial production point of view it barely seems like a drop in the bucket given how many cars/phones/planes we produce.

Each balloon would cover a circle, so it'd be more like 1,250 sq km. I think it would be optimistic to expect each balloon to cover exactly that area. They'd need to overlap just to eliminate gaps, and I'm sure there's a random element to their motion, so that would require them to overlap more.

4 million balloons and routers would cover earth. That's actually not a lot compared to how many cars and cellphones are manufactured every year. If each one is $100 dollars, the total cost is only 400 million - less than google fiber installed in kansas city ($1 billion).

Google is a business. They're only going to spend big money on a project if they expect to make a profit. Even if it might be possible in theory to provide coverage to the whole world via balloons, you'd have to be willing to lose a lot of money. From the standpoint of philanthropy, you would most likely get more benefit per dollar in other ways. For instance, the Gates Foundation has spent a lot in Africa on fighting diseases, and I'm sure lots more money could be spent on that.

> Google is a business. They're only going to spend big money on a project if they expect to make a profit.

People say this as if this was a law of physics. It's not. The company buildings will not spontaneously combust if they try to do something that is not calculated to make optimal short-term profit. They prove, time and again, that they care about awesome/worldchanging more than they care about the money. As someone pointed out in other thread, they even structured their stock options so that grumpy shareholders who care only about short-term profit can't get in the way.

I'll be the first to give them credit when they follow through on something.

On another site, which was also on HN (http://googleblog.blogspot.com/2013/06/introducing-project-l...), they are more forthcoming about the scale of the challenge and the likelihood of success. I like the tone of that page much more than this one.

Maybe they figure that each new person on the internet will have a good chance of being a Google costumer. Or maybe they'll charge for access in rich countries, but give it away in poor countries.

I think you're missing something - you don't need to be connected to the internet constantly in order for the internet to be very valuable to you. If you have access once a week for 30 minutes, that's a lot more than never. That's enough to send emails and keep in touch with the outside world in new ways.

If the balloons could be controlled accurately enough to provide internet access at a specific time each week, then you have a point. But if they fly over unpredictably, guided by the wind, then it's not going to be nearly as useful.

Utility is not binary. If they provide intermittent access to people who've never had it before, that's a huge deal.

The site linked here has a tone of triumphalism, but you can tell from other information about this project that Google is approaching this realistically. I like how Google's engineer said that they're expecting many more iterations of prototypes. No one in Africa is going to be grateful for a system that's a nonworking pile of junk. Whether or not it's worth rolling out will depend on how well it works technically, and that will become clear through testing.

Someone went to an event and posted what he learned here: https://news.ycombinator.com/item?id=5887330

It's good to see the hard information, because this site is all marketing.

I've always wanted a near-live satellite image for Google maps.

If this takes off that'll be possible. I could find a parking spot from the sky. :)

Loon has no such capability.



A: No.

It does not have it now, because they want to focus on the basics. Once that is done, adding a camera is not a big deal.

But I feel like that would make the media -- followed by average people -- go crazy. Consider the current NSA news combined with "Google puts hundreds of cameras in the sky".

> But I feel like that would make the media -- followed by average people -- go crazy

Yes, that will be a two uncomfortable days of online nerd rage.

A handful of police departments in the US already have constantly floating omni-directional camera drone balloons. The technology is only getting cheaper.

Huh ... don't look now, but Google is already putting hundreds of cameras in the sky, daily. Various aircraft, even some satellites, are devoted to Google imagery. Having a new fleet of balloon-based "craft" providing imagery could easily fit into Google's M.O.

And what happens when "Google puts thousands of cameras on our faces" ?

Well like if you didn't have a camera in your pocket for years. Or like cameras hidden in pens weren't available cheaply for years. Seriously, the time for people to be scared of others recording things is has already passed (mostly unnoticed).

It's different, since people will actually have to buy and use those cameras, the pushback will probably focus less on Google and more on the wearers.

So the government seeing your data is bad, but a private company filming your every move is good...

Look at it this way: if I wanted the government to put a CCTV camera in front of my house and snap photos once a year, it would mean I thought the best way to publish pictures of my house was to take tax dollars, enforce it through legislation, submit it to the whims of politicians, suffer the fruits of their corruption as they sell me out, and on and on.

Me? If I want to publish pictures of my house, I would likely use some sort of internet service to do so... therefore:

The laws around private companies photographing private property are both (1) very restrictive and (2) well known and well tested in court. Compare that to warrantless spying only reviewed by a secret court – and even then the government fails to (1) keep it secret, (2) abide by their own court's rulings, or (3) successfully apply the intel to the stated cause of terrorist threats.

Basically, there aren't really very many things the government does well at all.

When Google wants to send a car down my avenue or fly a balloon over my house, they'll be regulated by laws that literally follow the Streisand effect... in other words, if I don't want you to see my backyard, I might have to put up a roof.

It'll be uncomfortable, yes. And maybe Google isn't the best company to publish live video of an entire continent to the internet. But if someone else doesn't beat them to it, I imagine Google is going there. What's your contingency plan?

You act like that's a good thing.

I don't want some Google balloon watching me every step of my day.

Even less so if what it's seeing is published near-instantly

I hate this line of thinking. Nobody is watching you. Nobody is monitoring you. Even with PRISM, it's a massive amount of data, and unless you are on some watch list, nobody gives two shits about you specifically.

Google could just buy some of those awesome drones that get realtime multi-gigapixel video for a 15km radius.

Oh they could also use them like bombs and drop them on the "bad guys". They just need to make sure they explode with enough intensity, imagine a shit ton of these "cheap" balloons (economy of scale + ad-supported) just falling out of sky, and exploding everywhere lolz

Can someone explain why this is more cost effective than other alternatives? The need for specialized antenna's on the ground seems to be a negative.

Also, does Google just subsidize the entire thing? How is this paid for long term? How is it so much cheaper than the alternatives that not only can they roll it out to 5-6 billion people, but also allow people living on cents a day to purchase the service?

If it's very cheap, it could pay itself just with the legions of potential costumers (or eyeball pairs, to be cynical) it brings google.

Do you have some back of the envelope calculations of cost of this versus satellite or extending copper wire? I'm fairly ignorant on the subject, but would love to hear more.

No, I'm also fairly ignorant on the subject.

Have you tried to fathom rolling out the fiber program in Kansas City? The machinery, the man power, the political will?

This is a no brainer.

This is great question. In urban area it is extremly cheap to buy area for towers. What area can a tower cover with LTE? What is the theoretical limit with different technology?

Unless of course it is a trade secret, how wonderful would be if the engineers at Google would release more detailed diagrams of the balloons, circuits, and how it is supposed to work (the pump, antenna and everything, not the simplistic version). Maybe they are having some challenges which can be crowd sourced for ideas too. They have obviously done a lot of research and this idea can certainly spawn other awesome ideas or help other indie projects.

Project like these can be a really nice experiment on teaching and getting ideas from bored internet hackers.

Any news about other "moonshot" projects that Google[x] is working on besides self-driving cars and Loon?

I think Google Glass started out as a Google[x] project as well. I'd love to be a fly on the wall in their division.

Seconded. I think it would be just as interesting to see the projects that they've killed. They've said before that one of the goals of Google[x] is to fail fast. I imagine they've already accumulated quite the graveyard of crazy proposals.

"Google's mission is to organize the world's information and make it universally accessible and useful."

Very well in line with their mission statement. I love it because it's so much "ah screw it we'll just try this crazy sounding idea"

[I wrote a paper during my university time that was a thought experiment on the availability of cheap internet for everyone...time to dig it up :D]

They are testing it in New Zealand hinterland.

I wonder what implications this is going to hvae for NBN in Australia. They are spending tons of money to launch two satellites to provide access to remote towns. Google is showing they can do it for cheaper, possibly, and with Fiber they are showing gigabit FTH can be done cheaply...

I'm actually more concerned about what happens to the NBN if there's a change in government (looking more likely as each day passes).

Project Loon to bring NSA Spying to Everyone :)

While the geek in me acknowledges this is an awesome project and I'm really glad to see them giving it a try, the sarcastic cynic in me had the first thought that this was a great way to bring the the remaining 2/3 of the world under NSA surveillance.

A network in the sky. This is Skynet. This is literally Skynet.

Unfortunately, if more people start using the Internet the NSA database will inevitably grow. I wonder how long it'll take the whole human population is covered by the NSA.

Wasn't "NSA's free access to Google data" debunked, like, 3 days ago? Why people still keep talking like they give away all the data?

That's the problem with mass media. Not only a lie spreads faster than the truth, it's also the only thing remembered after the dust settles.

Has not be debunked in any way. As far as I understand it's far from clear what the NSA can actually do. Declaration of officials who are under oath to protect secrets is worth nothing.

I'm more excited about this than Google Glass.

In some regards, so am I, though this comment made me think that in some regards, Loon is a necessary component of Glass's success. There are huge areas of even just the U.S. which don't have cell/wireless internet coverage and Glass would be fairly useless, or significantly less useful, there currently.

Yes, but do you really think those rural areas are going to want to walk around lie cyber dorks? The corp I work for has already banned glass on premises.

Whether people in rural areas will or not, someone not from a rural area who does use Glass may want to travel outside of cell coverage areas and keep using it. Furthermore, far in the future it's fair to presume such devices would be miniaturized and likely become more acceptable.

My back-of-the-envelope first-approximation math suggests you'd need "only" about half a million of these things to cover the entire surface area of the Earth. To me, that seemed a surprisingly low number.

This could also be used to achieve low-latency network access in passenger aircraft over the oceans.

Personal opinion: sometimes Google is just simply awesome.

I really hope every word they've written in there, reflects some of what their top management feels at a high level.

Wondering how the routing is working.

If my neighbour and I are 30km apart, and covered by the same balloon, does our traffic get turned around on the balloon, rather than hoping to the base station and then back again? I have to assume yes.

If we are 30km apart and we are actually connected to two neighbouring balloons, will they route traffic between them, and again, not via the base station? Again, have to assume yes.

So do I get a single IP address, or do I keep changing when the balloons are overhead? I assume I get a single IP address, but my home router is actually going to have to do some intelligent next hop routing.

Mesh networking is a challenging area when the base stations are fixed, with weather effects and moving end users an issue. Even more challenging when your "base stations" are moving, weather effects are significantly greater, and you're super limited to the amount of onboard processing you can do.

Assume it has to be something like BATMAN. http://www.open-mesh.org/projects/open-mesh/wiki/BATMANConce...

Remember late last year, when there were rumors about Google being in talks to acquire Dish network?

I couldn't imagine that Google would actually use Dish's spectrum to build a traditional cellular network. Building all those towers is a huge investment and I didn't think that played to Google's strengths. But, assuming that Loon needs licensed spectrum, maybe Loon (or something like it) was the plan?

Well then, wireless internet escalated quickly. In all seriousness tho', this was unexpected. They said that it will be working with an antenna of sorts that you attach to your house? Fast forward 10 years and it will be an integrated wifi replacement in every machine - giving internet to you even in places where there's no electricity. That's the future I'd give my liver to see.

I'm not sure they've thought through the effects of creating that many more consumers and producers to compete with - especially when service industries, services that can be delivered at range, are becoming so dominant. This seems like the sort of thing that's going to force down the price of labour again.

I don't see the percentage for developed countries - seems maladaptive.

I find stuff like this awe-inspiring, particularly w/r/t how it could be used to protect internet access for not only those who are underprivileged but also for those being denied internet access by autocratic governments. Think back to Egypt, Libya, Syria - how different could those have gone if the authorities had been unable to shut down communication networks?

When you've taken over the internet, what's your next move? Make the internet bigger of course.

I for one welcome our new Google overlords.

I wonder if Google are working with the Serval guys in Australia on this. They are turning Android devices into wireless meshes and have been sending up balloons with device + high gain aerials for quite a while... With the same project motives in mind: disaster recovery (with ground and pole based setup, including voice calls and txt messages with higher QoS over the network), and extending coverage to less populated parts of the world. Australia is a huge place. Wireless is the only sane option for large large parts of the country, and towers, with power - you have to get it there somehow , are expensive. Yes solar is often an option (though it makes the base station in the middle of nowhere a very attractive target for thieves) and I'll be very interested to see how Google will solve that one (assume balloons will be semi covered in panels plus quite large batteries, and possibly small amounts of wind generation?).

Doh. Reading the FAQ. Many q's answered.

I'm curious as to the commercial viability of this. Will Google offer access for free or will you require a subscription? As a throwaway account mentioned, Space Data Corp. already offers this kind of service, but to large industrial companies, that can afford to pay a lot more than people in currently disconnected regions.

What kind of technology are they using to send the signal from the balloons to the ground? Though there were succesfull attempts to establish succesful WIFI links from the ground to stratoshpere(even 25 miles as I recall) this was achieved using precise DIRECT antennas (impossible when balloons are in constant movement)



Project Loon currently uses ISM bands (specifically 2.4 and 5.8 GHz bands) that are available for anyone to use.

Still doesn't explain how they achieve good quality link

Awesome. There are probably alot of things you can do with metrics on usage, availability, even mapping/weather and some fun camera views and more beyond just a line in the ground. So many more informational uses with it being airborne.

I love how google has so much money they can spend money on pet projects like these. No wonder they attract people with so many "impossible ideas" that sometimes turn out to be pet projects worth spending come money on.

If this could bring the INTERNET to places like China (where there is only INTRANET)!

Awesome idea!!! This will eventually provide alternative to the current telecom carriers.... In the long run they could provide the service free for Google Wallet users... They'll make profit on Wallet transaction...

Can anyone involved in this project shed some light on how altitude is controlled?

Could servers be put on these, too? Like that drone idea the PirateBay had?

Next up, Loon servers? I can't wait having my torrents served from the stratosphere. I believe TPB was already considering having drone-based servers - balloons just seem so much cheaper.

So the balloons just float on the air channels 20km up and presumably circulate around world? I can imagine problems coming up with some countries objecting to these being over their airspace.

So the question, then, is how high do you need to go before you're no longer in a country's airspace?

As far as I know there is no international agreement on this. Altitude can vary from 30Km to 160Km.

Most countries seem to claim somewhere between 70Km and 100Km as their airspace.

If someone has more details on this, it's very welcome.

This master's thesis from 2005 ("The Vertical Limit of State Sovereignty") has a fairly comprehensive account of the history and present state of international law in this area:


I'm familiar with that. And the guy proposes 22Km as the limit.

Besides, a lot has changed in the use of airspace and space in the last 8 years.

Yet, I don't know of any authoritative source for the actual current claims of most countries.

After the communications platform goes up, then the other payloads can hitch a ride.

Imaging getting real time video from the stratosphere to everywhere all the time. The weather data alone would be awesome. The traffic data would be great. The IR imagery catching forest fires or amazon jungle burning would be great. An almost real time collection for finding illegal logging or toxic waste dumping.

When there are clouds of these ballons going overhead, who will notice an extra one or two that happen to have militarized payloads.

On first reading this, I had to check the date to make sure it wasn't April Fools because I thought it sounded a little silly... The more I read it though, the cooler it sounds...

Plenty of pictures on their G+ page: https://plus.google.com/+ProjectLoon

There are quite a few technical challenges here (weight of cables, stability at 20 mi, erosion by much intense radiation etc). Details on one of the previous efforts which was considered revolutionary but never became main stream can be seen here: http://www.spacedata.net/news040108.htm.

In the first video, yeah having a freakin kid tell you why the world needs to be connected and explaining concepts that he/she is having problem even pronouncing does make me wanna support this lunatic loon project so google can whore out the rest of the planet with their ads. Can't stop imagining how useful this will be to the #nsa

Wasn't Page about to focus? Shutting down small, useful and used tools because Google should only do a few thngs, but all of them well?

So will this litter the countryside, and even worse, the oceans with electronic equipment incl. highly toxic batteries? will they maintain and retrieve every single one of those balloons? who pays for that?

They address this in one of the videos. They can direct the balloons to land at collection points for recycling etc. How they do that I have no idea - but it's definitely something they've considered in great detail.

I could easily rattle off a dozen Google products I'd see killed if it meant another project with the the potential impact of Loon.

I find it odd that you assume they're going to take these highly advanced routers and litter with them.

I share the same concerns - what about the batteries and other products that aren't as expensive as the routers?

(from http://www.google.com/loon/faq/#pilot)


A: We are taking several steps to ensure Project Loon does not negatively impact the environment:

- We’re working to guide all balloons to collection points upon descent, so we can reuse, recycle, or dispose of the parts responsibly

- When balloons do not make it to one of these collection points, we will be offering a reward for reporting the location of the balloon so we can collect it

- We can track where the balloons land and we have a team of people who focus on recovering the balloons

- We'll have a boat for recovering balloons that go into the sea close to New Zealand

- We've started looking into using biodegradable films for our balloon envelopes

- Our balloons’ electronics are entirely solar powered.

"Our balloons’ electronics are entirely solar powered." I wonder if that means there won't be batteries. But no batteries would mean no service at night. Also, no navigation at night.

Pulling them out of the deep ocean would likely be infeasible. They might be able to make them float, though, and get them when they wash up on a beach somewhere.

Edit: there are some answers here: https://news.ycombinator.com/item?id=5887330 They do have batteries.

It somehow reminds me of the fictional Aquinas Protocol:


It gave people free bandwidth in exchange for being a part of a global survelliance system. I hope Project Loon doesn't have such a dubious second nature.

Any details about bandwidth and latency? Where are the gateway stations out of the balloon network? In the local region?

Google has mentioned that they had this sort of internet infrastructure in mind, right? Awesome to see it move forward!

I've often thought something like this would be good for providing internet access to the people of North Korea.

Until they shot it down ;)

Although I think this initiative is nothing less than totally awesome, I wouldn't want Google as my ISP.

Sounds cool, but there is one issue: latency. I remember using the Internet in the mid-1980s, specifically using servers in Norway and other locations from my office in San Diego. Because of a few satellite hops, the latency was about 2 seconds. It made remote use of Emacs interesting :-)

Amazing idea, but then the rest of the world (non-americans) are still worried about the implications of this for PRISM. Getting the rest of the world online is cool, but then the NSA gets to have access to their data,that's not cool and I was beginning to like Google!

Yes, everyone in Africa is waiting for internet to use with their smart phones.

Evading the censorship is a better idea but I am pretty sure China wouldn't allow such a thing to fly over their country. Or at least make it suffer with network attacks or DDOS or something.

What is "loon" about this? We've been flying stuff outer space for over 50 years.

Is really that impractical to use stationary WiMax access points instead?

Last time I checked range was around 30 miles, couple of stations will blanket a good area without the hassle these balloons bring (you need to pick them and relaunch every couple of weeks, etc...).

Some benefits over stationary towers is that you can easily move the balloons based on different things like natural disasters, political unrest or even to just improve bandwidth under heavy load.

"Never mind about the NSA, here's this wacky project for the public good!"

Applications are open for YC Winter 2021

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact