Will Whitehorn talks a good game:
He foresees uses of the spaceship for science experiments, for example as an alternative to visiting the International Space Station or using unmanned flights for pharmaceuticals companies seeking to use microgravity to change particles.
Later, the aircraft could be used to launch small satellites or take other payloads into space, Whitehorn says. “We could put all of our server farms in space quite easily…”
…Eventually, he sees the possibility of transporting passengers to terrestrial destinations in spacecraft outside the atmosphere instead of by plane. He says a journey from Britain to Australia could be done in about 2-1/2 hours.
“That’s a 20-year horizon,” he said.
I’d take that a lot more seriously if he had liquid engines…
And of course, he never misses an opportunity to bad-mouth the competition:
Virgin is not the only non-governmental party trying to develop space travel in the private sphere, but Whitehorn is confident it will be the first to take passengers into space.
SpaceX, led by veteran Silicon Valley entrepreneur Elon Musk, is developing space-launch vehicles but they are not designed to carry passengers.
Well, yes, if you ignore the Dragon…
And of course, XCOR might beat them, though if they don’t get to a hundred klicks, the claim will be that they’re not in space, despite the stars, curvature of the earth, and minutes of weightlessness.
Dragon isn’t Flying, Falcon 9 isn’t flying. Falcon 1 isn’t in revenue service. The burn rate at SpaceX is horriffic, and, there is very little cash rolling in the door besides COTS.
Dragon isn’t Flying, Falcon 9 isn’t flying.
Neither is SpaceShipTwo.
Why do you insist on continuing to be the village idiot at this site?
It always amazes me how people you call idiots continue to make valid points that are “glossed” over in your own reportage.
Mr Simberg likes name-calling
In what way is it a “valid point”?
(Hint: the fact that it is true does not make it a “valid point” except to those unfamiliar with logic.)
WRT his dissing of SpaceX Mr. Whitehorn does appear to be engaged in a bit of an apples to oranges type of exercise. After all it is not very likely, to be charitable, that Virgin Galactic is going to put someone in orbit before Spacex. Suborbital is nice but hardly the same as ferrying a paying client to the ISS or hopefully a Bigelow facility.
WRT to the “glossed” valid point highlighted by Mr. Lee, irrespective of whether the money comes from NASA or Malaysian clients, I seem to recall that SpaceX maintains that it is cashflow positive, and significantly has actually got something to show for the money it has been burning, which is more than can be said for certain govt agencies in a similar line of work. It is reasonable to expect that even if it only ends up as another ornament off the Florida coast, which would be unfortunate, we should expect to see a F9 launch before the end of the year. I don’t think anyone is arguing strongly that we will be seeing launches of SS2 this year.
Why in the world would a company want to spend millions putting a server farm in space other than a desire to drive themselves into bankruptcy?
The server farm point really hit me as odd as well.
When I read the article, I thought the discussion of SpaceX was more ‘background’ from the reporter instead of Will’s comments directly.
I’m with Paulus and Tom. Quite apart from speed-of-light lag due to up/down-link distances – which are, in my opinion, plenty enough to render such a proposal a non-starter on their own – it strikes me that server farms are at the absolute opposite end of the sophistication and speed-of-obsolescence spectrum from the sorts of things we ought to be wishing to see in cislunar space. Even with single-flight lift, flight frequency and cost/lb.-on-orbit numbers of a Simberg-ian nature (if I may coin a phrase here), one would be lucky to get a serious-scale, solar-powered server farm on orbit and operational before the inexorable grinding of Moore’s Law reduced one’s investment to powder.
The ideal space infrastructure investment is one that is as low-tech as possible and whose purpose will be foreseeably served by suitably well-designed hardware just as well 50 years hence as when it is first put in place. Habitable volume – Bigelow’s project – is one such thing. Fuel depots – which needn’t be much more than big, dumb tanks – are another. Things that are indispensible and can be made as simple and dumb as possible are what we need the most of.
The server farm thing is the idea that 1. energy is free because solar powers denser (higher flux? is that a more accurate phrase?) and 2. server farms draw so much power off the grid that they contribute significantly to pollution from coal power plants and that might matter in a cap-and-trade system. Remember, the energy use isn’t just to power the chips and things; it’s also to power the AC units to cool the rooms where all those electronic space-heaters live.
It’s not for small farms, but big ones, like Google, MSN, Yahoo, etc. might benefit from it.
That’s an interesting point about obsolescent , Mr. Eagleson, but I’d also point out that speed-of-light transfer for 200-300 miles to space is still faster than slower-than-speed-of-light transfer 1000 from my home in Tucson to Silicon Valley where the Google servers might be. Yes, I know that Google has local hubs, but a lot of the data transfer now bounces off of satellites, anyway, especially for trans-oceanic signals. Remember that the alternative to 10 big server farms that are in low orbit is 1000 server farms around the globe close to major internet hubs.
That ended up being sort of confused. If there’s one global hub, it’s going to use a lot of satellite transmission to get over oceans. If Google has local mirrors of its index (which I think it does), it will have far more on the surface than it would in LEO, each some high expenses that the high-capital-costing orbiting stations won’t have.
It may not be a closed business case, but it’s a promising one, especially with a long-run eye on the moon, which is 20% silicon and a-bunch% iron and titanium.
OK, I think everybody missed it… Virgin is quite right, they will be the first (only?) company flying passengers.
Everybody else is over on this side of the pond, where we fly participants…
The Server Farm idea is silly. Servers need a lot of maintenance. If you put computers up in space you need a maintenance staff or you need some seriously reliable hardware. Also given the speed at which computers are getting obsolete, it seems kind of silly to park them in a very expensive location. There may be some niche applications for Space Solar Power, but, computers don’t strike me as the right application
The other probem with the ‘space server farms’ is heat. AC is a nontrivial cost for earth-based datacenters – getting rid of all the heat on one in orbit would be a huge problem, I’d think.
I worked as an administrator for a while, and most maintenance can be done over SSH. There’s not that much physical maintenance.
The transportation issues do suggest a system that grows through parallelization rather than upgrading parts. That and the maintenance issue suggests an architecture that has a lot of redundant parts that are reassembled through the OS, which boosts costs. And the increasing parallelization would require very different programming paradigms. And no one has yet brought up the fact that modern processor speeds are due to smaller transistors that are much more vulnerable to radiation-induced bit-flips.
But I guess the question is: why do you think Google is sponsoring the Google Lunar X Prize? It seems like the people with the most knowledge are looking at the numbers very carefully. Do you think your instant skepticism is more reliable than interested investors?
I try not to actively disbelieve things that I don’t have to care about. Real opportunities are found when they aren’t expected, not where they are. I wouldn’t invest in it unless I’d run the numbers and really understood whether it would work. But, I’m not going to fail to run the numbers when the idea seems to have so many credible people discussing it.
The energy might be free, but the initial investment costs are so completely unattractive that it’s not worth it. You might as well build your own dedicated green power plant. And, someone correct me if I’m wrong, but wouldn’t the cooling problem be even worse in space?
David,
I really like your sense of humor.
– Jim
P.S. In Wil’s defense, I even slip up and use passenger occasionally, and I drafted the law that codified “participant”.
My beef with the orbital server farm idea is bandwidth between Earth and the server farm. Open air transmission just isn’t keeping up with fiber optics. My take is that orbital server farms only makes sense for applications which use wireless transmission anyway. For example, caching and built-in server services for a global wireless network.
SpaceX is NOT developing a launch vehicle designed to carry passengers.
The vehicle designed to carry passengers is the Dragon, which is clearly NOT a launch vehicle.
The launch SYSTEM is designed to carry passengers, and other forms of upmass and downmass.
*/ pedantry off
Just so we know what we’re talking about, the cost of electricity and AC for a 10,000 sq.ft. datacenter is between a million and milion and a half dollars. Although it causes big headaches for owners of datacenters due to the cutthroat competition and low margins in their business, the amounts are just not there to support a computer that costs $10,000 per kilogram. As the power requirements of computers grow, this idea may be revisited, but its time is definitely not now, IMHO.
Actually, I’d love to think about issues that an orbiting system would introduce (obviously you want a belt of servers continuously streaming overhead in some kind of resonance orbit, to keep the communication delay low, for example). But I cannot see myself getting a comfortable lifestyle by solving the problem.
Well, OK, if people are going to take this seriously…
At there base, computers are tiny wafers of silicon. It is very simple to make a supercomputer from 2 decades ago that fits into a single chip – including a flash drive and wireless com. This would weigh next to nothing.
Looking forward to the next twenty years, it is quite conceivable that a useful amount of processing power might weigh 100 grams or so. So you can launch 100,000 of them in a single go.
I think where this idea comes from if the fact that computers (think chips, not boxen) actually have an extremely high value to mass ratio – and this ratio doubles every 18 months. Even right now, a computer CPU and flash drive assembled on a cheap board could cost far more than launching it would.
“We could put all of our server farms in space quite easily…”
Yes, but we couldn’t put their radiators and power supplies into space easily.
Me, I much prefer the idea of putting your servers into an ISO container parked at a high latitude and using airside economisation and Earth’s hundred billion tonnes of atmosphere as a thermal radiator.
“I know that Google has local hubs, but a lot of the data transfer now bounces off of satellites, anyway, especially for trans-oceanic signals.”
Only in Africa and the Middle East, and only because the submarine cables are owned by regional monopolies.
I think data centers in space are a none starter for now. In regards to cooling one could use the same idea for the James Webb space telescope. Build a large shroud around the data center, park it at a L point and let your heat dissipate into the blackness. Really though, it is not the heat that would be a problem but the cosmic ray issue. Cosmic rays affect the data on your electronics enough as it is even with the Earth’s magnetosphere helping. Look at the ISS and how their computers are fairing up there. They have had several unexpected lockups of their systems. A data center in space would need a fair amount of hardening to radiation.
We’ll see server farms in LEO if there is favorable enough tax or gambling law.
the fact that it is true does not make it a “valid point” except to those unfamiliar with logic.
It depends entirely on the context used Rand. in this context it was a valid point. Thanks for playing though.
Re: Server Farms.
Sam is right, they have a place if they can operate outside national laws, but that’s about it.
Johnathon, the problem isn’t the 200-300 miles of speed of light connection to LEO, the problem is where your current server farm is in the 20,000 odd miles of LEO at the time. Unless you’ve multiple farms which sync continuously between themselves using a network of satellites.
Either way, the whole system is likely to be much slower for most people than fiber.
Josh makes the other point – there’s a lot of high energy crud up there that isn’t all that good for long term computer use.
It depends entirely on the context used Rand. in this context it was a valid point.
Only to an illogician.
The Server Farm idea is silly.
Someone call SETI. I’ve just received signs of intelligent life from Jack Lee.