Ok, lets start with the assumption that no existing country will be able to provide the kind of guarantees that a real data haven needs to operate -- the costs are, in the general case, way too high. However, it should be possible to build a nation out of the sea in a remote corner of the ocean that would be able to provide those guarantees, since it would be built for that purpose. Let's review what I consider to be important criteria for an appropriate location:

  • A sea mount would be an ideal location, as it would provide a relatively shallow underwater foundation, and is unlikely to be currently considered important by anyone, easing costs of acquisition.
  • This sea mount should be outside the normal territorial waters or economic exclusion zone (200 mi) of any sovereign nation.
  • It should be either unclaimed or claimed by a relatively poor nation, so it can be cheaply bought.
  • It should be in relatively calm waters

My personal favorite is a location called Minerva Reef, several hundred miles south of Fiji, and claimed by same. Fiji maintains an automated lighthouse station there, but nothing more. It is also a shallow coral reef -- the top breaks the water about half of the time, but is awash at high tide. This should make construction relatively easy for a remote sea mount. It is also out of the main hurricane paths in the area. Fiji is relatively poor and might be willing to sell for a reasonable price, especially if the builders promised to build the platforms in Fiji before towing them out to the reef, and employ Fijian workers on site. A large platform, possibly including a short landing strip, should be doable for under $10 million dollars.

Small note here, there would be a great deal of outrage at the simple idea of building anything on a reef, let alone a data haven.

I'm not saying it's a bad idea just that building on a reef wouldn't be the best choice because of the enivronmental problems

I strongly believe the best place for a data haven would be inside an orbiting satellite. With the size and cost of storage going down every year, it should be feasible to stick a fair amount of gigs into a payload loftable by Ariane 4 (the -5 blows up too often) or a competitor like Sea Launch. In any case, the advantages of this are manyfold:

  • There's already a communications infrastructure for talking to a satellite. Building one on a reef might cause trouble.
  • Sovereignty is definitely less of a problem up there.
  • It's much harder for someone to physically try to steal the data, especially without being seen.
  • Power is free up there, if you have large enough solar panels.
  • Cooling can be a problem, but passive radiator fins work, and the temperature differential is free, too.
  • If you set up a secure enough communications link (freq, protocol, comm. times, codes etc. known only to you) you can be pretty darn sure that nobody will get to it.
  • If you really, really really wanted to be secure and paranoid, you could use a quantum encryption system to communicate with the satellite. These systems are best devised using lasers, already a viable way to talk to your datacenter without the hassle of fiberoptics. You'd know if someone tapped your communications that way. Also, you could hardwire the satellite to always broadcast a copy of anything it sends out, and send the copy directly to you. Then, you could compare logs and know if someone had managed to nab control of the bird.
  • A paranoid self-destruct option would be quite easy to implement, requiring only a small retrograde thruster and perhaps a small charge to sever the main antenna.
In any case, the real issue would be funding this...and then launching from somewhere that didn't legally 'flag' the spacecraft as its territory. Sea Launch would be ideal for this; as a standard operating procedure they take their payloads out to the middle of the ocean (near the Equator, for maximum efficiency) and launch there. If you were really paranoid, you might deliver the payload to them at sea.

Actually, this is a fairly cool legal and technological problem.

Log in or register to write something here or to contact authors.