Hacker News new | comments | show | ask | jobs | submit login
Google Loon balloon crashes near Yakima (seattletimes.com)
97 points by gregholmberg on June 3, 2014 | hide | past | web | favorite | 45 comments

In the 90s, I would sometimes trip over ethernet cables that had been strung between server racks, across an aisle.

Will this be the decade where I get to experience being struck by falling routers and switches, far from any data center?

(EDIT: Simply remarking on novelty of problem and upward growth of network. No criticism of any firm implied or intended.)

I took down an AS/400 server once by tripping over two Type 1 token ring cables[1] that had been plugged together in the middle of a walkway. No-one ever owned up to that setup.

[1] http://en.wikipedia.org/wiki/File:IBM_hermaphroditic_connect...

My hands down favorite Dilbert comic ever was a token ring joke:


Given that it was token ring that might have isolated the AS/400 but the MAU would have healed the ring and everything would have been fine.

Oh god, I’d blocked from my mind the pain of ring dropping. We had MAUs for the servers, but for some reason ran the rest of the office on CAT-5 TR gear from Madge without any MAUs. Just took someone to unplug and that section of the office would lose network connectivity.


I used to work for Madge. Why were you running without MAUs?

Possibly the more important question is ‘Why were you running a token ring network in 1999?’! I have no idea why our network was structured like that, this was my first job in the industry and I was a lowly PC tech at the time. If I had to guess, the word that springs to mind is ‘legacy’.

>'Possibly the more important question is ‘Why were you running a token ring network in 1999?’!'

I've worked on a token ring network as recently as 2006. Thankfully, part of the project was migrating off of it.

Eh, I was a user on a token ring university network almost that late, and it was superior to the congestion-crippled ethernet networks at more 'advanced' institutions. That said, I definitely didn't enjoy all the practice I got in building token ring-enabled Linux kernels.

>Will this be the decade where I get to experience being struck by falling routers and switches,

especially if the falling equipment identifies you as the softest landing spot in the vicinity (thus providing for lesser damage to / higher chances of the equipment survival :)

> far from any data center?

until off course it is the air borne (near-space) datacenter itself.

The children's story will have to be revised.

Chicken little decided to go for a walk outside and was killed when the internet fell on him. The sky continued unabated.

The word 'crashes' could easily be substituted with 'drifted into'.

Dropping at 2000 feet per minute is not quite free fall, but it's a long way from a drifting dandelion seed. I wonder if the balloon payload left a crater.

At 22 MPH? Probably not considering it has to be light enough to be a balloon payload in the first place.

I've launched a handful of latex weather balloons to near-space. They all descend on a parachute and have a target impact velocity of ~2000 feet per minute. None have made so much as a dent in the earth so far.

I believe you. :)

I have no experience with balloons at all, but even a plowed field feels like a sidewalk to me when I run into it at the same speed under canopy. I leave a sizable mark, and I have seen friends break femurs, pelvises, etc at the end of a similar descent.

I grew up outside of Yakima. Northern/Eastern Washington State has some very desolate places, maybe that area will be the actual user testing ground too.

There is something almost poetic in this. The perils of accidental infrastructural intercourse.

What do the Loon components look like up close? Here's a gallery showing the solar panels, avionics bay with covers open, and some staffers.


I would still go for cables on the ground ... or underground.

There are definitely advantages to being completely wireless, building permits being one of them.

I like that the article referred to it as a trial balloon.

Time for Bale Breaker Field 41 :)

Guess that's a hazzard of balloons all over the place. Hope one does not take out a jet.

Balloons are launched all the time, Google's balloons are no different than the many many weather balloons, etc everyday. Their regulations are pretty easy [1] to find and follow. Basically, you call them, let them know when and where it's going up, when and where it's going down, and they'll put out a notice that pilots look at letting them know that there's a balloon out there they should look out for. If it gets loose, you call them again, say where you think it's going, etc, and they'll update the notifications accordingly. Stuff like this has been handled for a very long time.

[1] http://www.chem.hawaii.edu/uham/part101.html

Indeed they are launched all the time. But, we are talking about the case that the practice becomes popular and the skies become a dense minefield.

The cruising altitude of Google Loon's is well above the cruising altitude of any commercial jet. This accident was caused by a balloon coming down, a much rarer occurrence.

I don't expect to hear a postmortem from Google, but I'd be astonished if this wasn't a malfunction of some sort-- These balloons almost certainly have an emergency cut-down device of some sort capable of safely and rapidly returning the payload to Earth.

That's an interesting back of the envelope question to work out. Something with volume occupied by all aircraft aloft at any given moment, volume of all balloons at any given moment, the total volume of the shell in which commercial air traffic takes place.

I think we can safely forget about the chances of spyplanes hitting balloons, the volume of space versus the number of spyplanes would make that a non-issue, even if there were a lot more balloons.

So 35000 feet (11 km give or take) would be a reasonable upper limit. Let's assume the worst and start from 0, you have a shell above the earths surface up to 11 km above it, which has an approximate volume of: 510.1 million km x 11 = 5610 million cubic kilometers.

That's a lot of space. Every cubic kilometer is 10^9 cubic meters, so 5.6x10^18 cubic meters.

I don't know how many aircraft are typically aloft, but let's say it's 20,000 craft and they're all of the very largest variety (say A380, or Boeing dreamliner). They're approximately 60 meters long, and 6 meter in diameter, so that's 1700 cubic meters, let's double that to include the wing volume, so 3400 cubic meters.

We have 20,000 of them, they're all aloft at the same time, so all the planes take up approximately 68,000,000 cubic meters.

Now for the balloons, they're 10 meters in diameter, worst case they are 50 meters high or so (instrument package dangling below the balloon, assuming a cylinder with a radius of 5 meters and a height of 50), so about 4000 cubic meters. ('assume a spherical cow of uniform density').

So how big is the chance that one balloon intersects in all of space with the volume of all the aircraft given that both have all of the atmosphere to play cat and mouse in?

68,000,000 / (5.6x10^18) = 0.000000000012 (the chance that any given cubic meter is part of the space occupied by an aircraft) multiplied by 4000 (the number of cubic meters in a balloon) is about 0.000000048. So that's pretty small but non-zero, multiply by the number of balloons aloft at any given time, but keep in mind that most of the factors here were taken very pessimistic (as in, favouring the collision). The calculation also totally ignores the relative speeds of the two types of vehicles, ascent speed of balloons, the time factor, ability to manoeuvre and so on.

See also:


edit: extensively edited after avoid3d spotted a crucial error in the math, I had dropped 6 orders of magnitude.

The problem is that jetliners don't have all that space you calculated to play around in. They actually have a few well defined corridors and altitudes in which they can operate - the straight lines drawn from one VOR (effectively equivalent to a major airport) to another at 1,000 foot intervals.

Detours are costly due to time and coordination (air traffic control, other aircraft), and reacting to seeing a balloon and moving the aircraft isn't that easy when you're traveling at 300+ MPH in an aircraft which turns like a cargo ship. And that's assuming you can even see the balloon in time to react in the first place.

And that's just the commercial jetliners. Private jets go higher and faster (about 50,000' and 700mph), while GA aircraft fill the skys below 14,000'.

Granted, this still leaves a lot of room in between these major aircraft corridors, but if a balloon should ever intersect with one of them, it's going to cause havoc, even if there's never an actual balloon/aircraft incident.

> They actually have a few well defined corridors and altitudes in which they can operate - the straight lines drawn from one VOR (effectively equivalent to a major airport) to another at 1,000 foot intervals.

Until NextGen (ADS, etc) finishes its rollout and everyone flies direct instead of on Victor airways and VOR to VOR.

True, but even then there will still result in "air highways" from airport to airport.

I would agree with this. Still, the number of "air highways" would increase drastically.

You lost 6 orders of magnitude to this IMHO:

5610 million km^3 = 5.61 * 10^3 * 10^6 km^3 = 5.61 * 10^9 km^3 = 5.61 * 10^9 * 10^9 m^3 = 5.61 * 10^18 m^3

But you said the resulting volume was = 5.61 * 10^12 m^3.

I think you forgot that it was 5610 million km^3.

Ah! I knew there was something terribly wrong there somewhere (and probably that's not the only thing). Thank you very much, I'll edit the comment to fix it.

Keep in mind that balloons are much cheaper than planes to build and operate. Worse yet, the populated areas on earth are only a fraction of the whole surface area. Expect balloons to make dense clusters rather than evenly distributed over the sky. Needless to say, there will be more flights to more populated areas as well. Increasing the likelihood of collision to a level that the pilots will need to constantly monitor the skies will increase fatigue and cause indirect problems. The cost of all this will be felt by the aviation industry, and will be passed on to us.

The chances of hitting a jet are remote, the balloon would have to loiter at the right altitude for that chance to increase appreciably. Most will go higher, pop and then fall.

The barrage balloons of WW-II had to be anchored very carefully to avoid having them go into the stratosphere. As the balloon expands (which it does when it goes higher) it will become more buoyant, not less so there is a positive feedback loop in there which usually ends in destruction unless you take precautions. Such anchoring requires very long cables, which makes them a bad choice to defend against jets.

So, the risks are non-zero and if one were to get sucked into a jet engine (especially the payload portion) the mayhem would be considerable, but they are so small that a 'notice to airmen' suffices unless you're operating very close to an airfield when you launch.

What is interesting about this incident is how far the balloon came down from where it was launched, it must have travelled for a long time, maybe even circumnavigated the globe more than once before landing.

Are you sure about positive feedback? Balloon does expand, but air outside becomes thinner, so buoyancy will actually stay constant until the walls are completely unfold, and then drop as gas inside will expand less than necessary to maintain buoyancy.

The gas inside will want to expand further, but it will be restrained by the envelope. Imagine the outside air pressing inward until there is an equilibrium. If the balloon then rises again (for instance, because it warms up) then it can get to a point where the outward force on the envelope exceeds the capacity of the envelop material to restrain that force.

For flexible envelopes this is more or less inevitable (balloon rises->atmospheric pressure drops->balloon expands->density decreases->balloon rise etc), for more rigid envelopes it is a balance that may work out in favor of the balloon staying in one piece (oscillating in altitude as it cools down/heats up again with the day/night cycle), or it may burst depending on the pressure differential. Most of them are pretty flimsy.

From: http://www.stratostar.net/faq.html

"When a balloon is filled on the ground with lift gas (helium or hydrogen), it can range in size from 2.5 ft to 8 ft in diameter. During the balloon's flight it will grow more than 4 times the diameter and upto 83 times the volume measured at launch, until it can't strech any more and will burst! A high-altitude weather balloon filled with 268 cu/ft of helium will have a diameter of about 8 ft at sea level, but as the balloon climbs through the atmosphere it will expand to 35ft in diameter and will have a volume of 22,449 cu/ft before it pops."

My point was it's not becoming more buoyant, i.e. lift can only drop when it ascends. I wouldn't argue that it can burst, but it depends on material. As an edge case it won't go to the Moon as lift will be 0 in the vacuum. So unless there is a limit on balloon size the lift will be constant, otherwise it will drop during ascend.

It won't go to the moon because it does not have sufficient velocity to escape the gravity well. Not because there is no lift in a vacuum (it will never reach a vacuum).

It will simply float to the top of the atmosphere (if it stays in one piece) to the point where the weight of the baloon is balanced by the buoyancy. Just like a rubber ball will not float 'on top' of the water that supports it but slightly inside it.

It's Archimedes' law applied to balloons.

Jet aircraft are a little more manoeuvrable than power lines.

At the speed your average jet is travelling and the fact that approximately half the traffic is at night I'd hope for a radar transponder on the balloon (or at least some kind of reflector) because I highly doubt you'd be able to respond in time when going on visual cues alone, you'd need at least a few kilometers warning.

They have ADS-B transponders. You see them sometimes on flightradar24 floating around at FL650.

> The helium balloons are all “floating” around 1,000 feet to the southeast of New Zealand

I can't make soup of that.

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