[cryptography] How big a speedup through storage?
admin at pilobilus.net
Tue Jun 24 23:26:29 EDT 2014
> Date: Sat, 21 Jun 2014 03:10:58 -0400
> From: grarpamp <grarpamp at gmail.com>
> To: Crypto discussion list <cryptography at randombit.net>
> Cc: cypherpunks at cpunks.org, cryptography <cryptography at metzdowd.com>
> Subject: Re: [cryptography] How big a speedup through storage?
> <CAD2Ti2_Dcrqrnp30QJJdoH95zCF2E7QH1nw_Lrn9v-0nv5opow at mail.gmail.com>
> Content-Type: text/plain; charset=UTF-8
> I meant as answer 'estimates on how big' question. Take what we know
> about storage, figure in some good efficiencies for the 'storage only' case.
> And figure what can be bought and operated year on year per foot. You could
> hide/support $1B + $1B/year but $10B/yr would be hard given entire intel
> budget is $80B, or $50+B if you drop mil. So...
> 1) How big can you get within budget?
> 2) What can you do with it re: a) crypto, or b) otherwise?
In re the Bluffdale facility, it has about 100,000 square feet of
actual data warehouse, and is equipped to support a 65 megawatt
power demand and uses 1.7 million gallons of water per day for
cooling. This gives a baseline for estimating the power density and
processing capacity of the largest (known) U.S. TLA data center. An
open question: Is all that space for long term storage enabling
retroactive surveillance of future enemies of the State, or is some
portion of its capacity dedicated to cryptography?
More information about the cryptography