Google’s Balloon Internet Experiment, One Year Later

http://www.wired.com/2014/06/google-balloons-year-later

by jervisfm • 12 years ago

212 61 12 years ago

wired.com

tgb• 12 years ago

It's always great to find that my initial pessimism was too hasty! Best of luck to this exciting project.

lchengify• 12 years ago

Reminds me of the low-orbit satellite network in Society of the Mind.

IIRC, the protagonist wanted a internet-connected TV product with a small antenna. The only way to do it was to create a satellite network where the satellites were so close they would re-enter after a short timespan. It was countered by a continuous stream of launching satellites.

Replace "TV" with "LTE Phone", it's basically the same idea / value add.

Wingman4l7• 12 years ago

Surely some variation on high-altitude aircraft would be more cost-effective than continually launching satellites?

taterbase• 12 years ago

The http://internet.org project is doing just that.

https://www.youtube.com/watch?v=pxX6r-xDgG4

bostan• 12 years ago

Chinese users are excluded, and if not, HN numbers will be way higher.

Gustomaximus2• 12 years ago

Does anyone know how retrieval is ideally planned opposed to the current search party option?

There must be a reasonable cost for each balloons electronics, as well as the risk of this equipment hitting someone with balloon/time increases if left to random descent.

agildehaus• 12 years ago

They've already had one crash into a power line.

johansch• 12 years ago

The other option they are working on besides balloons seems more manageable in that regard:

http://www.cnet.com/news/google-buys-solar-powered-drone-com...

http://motherboard.vice.com/read/google-will-beam-gigabit-in...

micampe3• 12 years ago

There is one thing I don’t understand: the balloon can have a powerful enough antenna to send information to a phone 60,000 or more feet away, how does the phone talk back to the balloon?

jonknee• 12 years ago

Cell phones already do that now and the sky usually has the best line of sight.

Gustomaximus2• 12 years ago

It seems there is a ground based repeater, at least initially. See the second article picture: 'Project Loon team members install a Loon Internet antenna while schoolchildren look on.'

That said we have satellite phones now. Given the lower altitude of balloons getting a signal back should be realistic over future developments.

micampe3• 12 years ago

The article says the repeater is only for wifi, phones connect directly.

sjtrny2• 12 years ago

It's not direct phone to balloon communication. It's phone to fixed ground based wifi, fixed ground base to balloon. Although this is kind of glossed over in a lot of PR.

magicalist• 12 years ago

from the article:

Google made a different kind of advance with Loon when it added the capability to send data using the LTE spectrum—making it possible for people to connect directly to the Internet with their mobile phones. (Loon’s original Wi-Fi connection required a base station and a special antenna.)

gonzo3• 12 years ago

Antenna gain is retro directive.

micampe3• 12 years ago

What does that mean?

PanMan• 12 years ago

GSM cell's can be up to 40 km after which the problem becomes the timing (and 70 with extended cells). With cell's overhead, 20 km should be oke.

ChuckFrank• 12 years ago

I was invited to apply to Project Loon, which I did and unfortunately I never hear back from them. I wish I had.

darklrd• 12 years ago

It seemed like not feasible when I heard about it. The way they have approached every obstacle along with extensive testing - it's commendable.

howon92• 12 years ago

I think they deserve to be called more than "it's commendable." I think they did awesome!

zw1234563• 12 years ago

I think it is really great that Google is trying this, I don't want to be too pessimistic either as they seem to always defy the odds, but one of the problems with any type of high altitude platform whether it is LEO, Geosynchronous balloons or otherwise is that they cover a huge area, just by nature, which limits the capacity. Terrestrial cellular technologies get around this issue by being low to the ground so that the RF energy is dissipated and you are able to re-use the frequencies with the next cell over. With such high altitudes that is just not feasible. So I think this approach will always have limited capacity. Also, the fact that it is moving will be a challenge since both LTE and WiFi use OFDMA which is very sensitive to Center frequency offset that can be caused by the Doppler shift as the balloon moves. At 60kft, the uplink is going to be a challenge since most handsets transmit at 200mw max so it will be difficult to achieve high data rates in both directions. There are a number of other technical challenges that are ahead of them as well, but I am very impressed with their perseverance. I hope they are successful.

gonzo3• 12 years ago

Doppler shift isn't a huge problem for them. They use SDR and know their velocity.

zw1234563• 12 years ago

The velocity is dependent on the vector, which you do not know since it is dependent on where the user is. If the ground station is a known location perhaps, but otherwise you cannot predict the vector. For commercial LTE at 700Mhz, the offset is 5Hz, above that the error rate starts to exceed the specs. It is pretty tight for OFDM.

frozenport3• 12 years ago

If you want to get the next billion connect why balloons? Why not a conventional grid? Why not improve internet service in Mumbai? Is there a more direct way to get the next billion connected? Does Google think there is no competition in these rural areas?

Goopplesoft• 12 years ago

Because if you get this unconventional method right your barrier to scaling reduces in the long run.

eru• 12 years ago

The balloons are one project by one company. Google has more irons in the fire, and there are more companies.

Let a thousand flowers bloom.

Maakuth• 12 years ago

At least balloons are impossible to steal or sabotage, which might be a benefit for areas where no functional state is present (try Somalia). I can also imagine they are very cost efficient and fast to deploy to areas where connectivity is needed or is rapidly lost, say after an earthquake.

BCM43• 12 years ago

Really? I'm curious how hard it would be to shoot one down.

Steko• 12 years ago

It's hard to take this next 5 billion stuff seriously when there's still no P2P mesh support built into Android.

contingencies4• 12 years ago

You absolutely hit the nail on the head with this comment.

Furthermore, does every developing country want high bandwidth surveillance by developed countries? Because that's what a digitally unsophisticated populace and a crap-ton of smartphones with social networks, high resolution CCDs and IP telephony with dot gov access holes (eg. Skype) does...

Personally I think the China model (limit international bandwidth extensively, block or muck with traffic on larger foreign surveillance systems slash free web based services with invite buttons and 20-second polling apps with GPS requirements) actually has some decent properties in terms of decentralizing global power when viewed in this light: "ZOMG, furry foreign kittens!" They're like, exactly the same in China.

NortySpock• 12 years ago

I was also expecting them to implement bundle routing / delay-tolerant networking. It just makes so much sense for mobile devices with intermittent connectivity.

ihsw• 12 years ago

Because then you have to tear up the ground and maintain the grid's quality and physical security, this would involve managing relationships with numerous jurisdictions on many levels.

Also the bandwidth/latency needs are initially going to be quite a bit lower, which eases capacity planning.

johnsteve• 12 years ago

Good initiate taken by Google and the other company.

calinet63• 12 years ago

The most striking part of this is the systematic quality improvement they implemented. They went from an expert saying that long-term balloons were impossible, to making it a reality by understanding the true engineering variation causing failures.

We should look at every endeavor in this manner to improve quality. Look up W. Edwards Deming, wonderful stuff on quality control that can be applied to software just as well as manufacturing.

mathattack3• 12 years ago

Deming's career arc has been interesting. Neglected in the US to revered in the Japan to revered in the US to forgotten in the US. Maybe now that we don't fear Japan so much, we don't feel the need to obsess about his views on quality.

Many of his ideas stand the test of time. He just doesn't fit neatly into any academic curriculum.

calinet63• 12 years ago

Such is the flaw of academic curricula. Extreme specialization in a world of complexity.

Quality in Deming's point of view requires a deep multidisciplinary understanding: statistics, psychology, engineering, business, economics, and more.

The true irony is that now, more than ever, coming out of a recession and having a corporate and manufacturing issue in the US, we need the manufacturing and quality prowess that Japan exhibited post-WWII. We need Deming.

I believe his main problem was that he was a horrible writer. We need someone to take his ideas and clarify them much better than he could.

metaobject• 12 years ago

The article stated that Titan may be able to use Loon's wind data (exactly for what it did not say) which implies that Google will (obviously) be collecting and storing wind data for each balloon. This may open up a new line of business for Google. We could soon see them selling access to their data to the government (wind data, that is) or to other corporations that are invoked in similar activities in that part of the atmosphere.

Furthermore, using data collected by the balloons, in conjunction with model forecasts generated by NOAA, could lead to some interesting/challenging/fun new algorithm development - using wind data from some subset of balloons to predict the trajectory of some other subset of balloons. Sound like fun!

Shivetya• 12 years ago

Is there a reason not to anchor some of these? I know your very height restricted when anchored because of line weight, but why not anchored balloons as poor man towers? Or are they simply trying get beyond most weather?

jackgavigan• 12 years ago

..since NOAA only supplies forecasts for 16 days out, Google now has to make sophisticated guesses using a giant database of historical wind and weather data.

It would be interesting if, as a side-effect of Project Loom, Google ended up becoming the world's leading weather forecaster.

socrates1998• 12 years ago

I am curious, why can't you tie the balloons to the ground and just have them stay in the same place permanently until they lose their air?

Could you then create a continual service over a grid like network of these balloons? Or are the balloons too expensive to do this?

teraflop• 12 years ago

The balloons only have so much buoyancy, and 60,000 feet of cable would end up being pretty heavy. Not to mention that the cables would be a hazard to aircraft navigation.

juice13• 12 years ago

And people could steal the equipment by pulling the baloon down.

computer• 12 years ago

You could always make the cable out of balloons. But I guess that would be rather expensive.

Sonthun• 12 years ago

Has anybody seen any specs on these? Weight? Number of batteries? Solar panels?

lafar6502• 12 years ago

What about all kinds of national radio spectrum regulations or broadcasting/telecom laws? Will these baloons be allowed to operate everywhere?

metabren• 12 years ago

Am I the only one that thinks these are creepy?

a full ring of between 300 to 400 balloons circling the globe

I don't want Google looking back at me whenever I look up into the sky.

tomcam• 12 years ago

Article does not address danger to aircraft.

Google’s Balloon Internet Experiment, One Year Later

Related Stories