[FoRK] Server-sky and beamed energy / lighter than air / shuttle-glide transport
Stephen D. Williams
sdw at lig.net
Wed Jun 16 13:46:29 PDT 2010
On 6/16/10 9:52 AM, Aaron Burt wrote:
>> Data centers passing overhead in the sky make a lot of sense,
>> if you think of it.
> We're on it: http://server-sky.com/
Cool. Except that you are cluttering orbit which is a potential
nightmare and has potentially astronomical cost at some point. Build in
a failsafe apoptosis mechanism and you may address it well enough: Lose
radio / computer control -> aim for burn up in the atmosphere with a
reserve booster while blaring a warning beacon.
And bandwidth to the ground will be weak compared to fiber, although
multi-channel multi-spot multi-polarity phased array wideband
transmission can help a lot.
Otherwise, cool. Set up LEO (for anytime easy radio reach of handheld
devices), geosync, and far orbit, plus a heavy station on the moon and
we'll have a place for Wikipedia, Wayback, and Facebook that will
survive peak oil, or whatever.
It seems that microwaving excess energy to the ground wouldn't be such a
good idea overall, but it occurs to me that beaming it to
lighter-than-air stations / transports / floating servers might be
perfect: Safe to people on the ground, much less atmosphere to get
You could have a long-haul transportation system of high-altitude
floating ships with detachable cargo / cruising modules. Just fly up
the outgoing module, dock it first, then undock the incoming modules and
glide them to a landing strip. Lighter than air support for carrying, a
combination of jet stream / solar / microwave-beamed-from-orbit, and
some auxiliary fuel for maneuvering or stabilizing. The modules could
be shuttle-like auto-land systems in case of emergency or just to be
able to drop to the desired location without active ground support.
I don't see any obvious scalability issues. Not sure about feasibility
of continuous stable around-the-globe jet-stream following, seems like
the flow is fairly variable even just in the US. The degree of energy
input needed will depend on need for speed and predictable path vs.
efficiency. Significant mass will be expensive to reorient, so using
the jet stream only when convenient might be best.
> If that doesn't work, add "datacentre capacity" to your "Peak X" scenario.
> As much as I want to work on large-scale computational efficiency,
> datacentres are just as subject to the efficiency paradox (increased
> resource efficiency leads to more resource demand) as any other resource.
More information about the FoRK