On Wed, Jul 21, 2004 at 10:20:36AM -0400, Tyler Durden wrote:
Yes, but I think it's fairly clear that if one needs to dissasemble the OC-Ns in the field, you simply need too much gear. It's going to be far
It's clearly not viable to process much underwater. How much machine room square meters do you need at those cable landings, though? http://cryptome.quintessenz.at/mirror/cable-eyeball.htm
easier to grab whole swathes of it and ship it back to Montana or wherever for it to be sifted through later.
There is no "later", there's only "elsewhere". Traffic filtering is an embarrassingly parallel problem. It's the data mining that needs to integrate and correlate. Here is your centralized bottleneck. How many .gov in http://top500.org/list/2004/06/ ? Data mining is different from Linpack.
What they probably do, however, is grab specific DS1s/3s locall and switch those via CALEA back to optical access points, where all of this stuff is pulled together into OC-192s or (very soon) OC-768s. As Variola suggests, once you get it back then you can plow through it at your leisure. Got a disident you want to shut down? "Surely he's said SOMETHING over the last 2 years that you could incriminate him on....find it, dammit!"
-- Eugen* Leitl <a href="http://leitl.org">leitl</a> ______________________________________________________________ ICBM: 48.07078, 11.61144 http://www.leitl.org 8B29F6BE: 099D 78BA 2FD3 B014 B08A 7779 75B0 2443 8B29 F6BE http://moleculardevices.org http://nanomachines.net [demime 1.01d removed an attachment of type application/pgp-signature]