Help with my secret science project

[Blue]

So I've been doing research and stabs at a project off and on for about 10 years now.

I have a new theory and I am starting from scratch with it, and a new set of algorithms.

Anyone know anything about large-scale databases?

I'm going to need a way to store very very large amounts of binary data (starting with  a few Tb, and going up from there) with scalable inter-connections (each database entry/line will need to be able to associate with a theoretically infinite other database entries or lines).

I would use the tried and true SQL, but I'm not entirely certain that it would be the fastest way to go about this considering the scale. Using Google's App engine or other third-party datahosting would be too costly, so I'll be building another cluster network to house the new project. So whatever database method is used, it must be able to load-balance fluidly over a cluster of machines.

Anyone have any suggestions?
Methods of storing and retrieving high speed binary data in seemlingly random access?

[AstronautClock]

get your own hosting company

[Blue]

get your own hosting company

hosting is not a very lucrative business, especially if you are far from national internet backbones.