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.<p>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.
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.
<i>> ..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.</i><p>It would be interesting if, as a side-effect of Project Loom, Google ended up becoming the world's leading weather forecaster.
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 (<i>wind</i> data, that is) or to other corporations that are invoked in similar activities in that part of the atmosphere.<p>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!
Does anyone know how retrieval is ideally planned opposed to the current search party option?<p>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.
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?
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?
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?
Reminds me of the low-orbit satellite network in Society of the Mind.<p>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.<p>Replace "TV" with "LTE Phone", it's basically the same idea / value add.
Am I the only one that thinks these are creepy?<p>> a full ring of between 300 to 400 balloons circling the globe<p>I don't want Google looking back at me whenever I look up into the sky.
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?<p>Could you then create a continual service over a grid like network of these balloons? Or are the balloons too expensive to do this?