Re: Anon

[Adam Shostack <adam@homeport.org>]

Dale Thorn wrote:
| Adam Shostack wrote:
| > Brin's argument has two ideas that I find annoying.  One is that the
| > changes he forsees are inevitable, the other is that security
| > is not about economics.
| > The idea that universal surveillance is inevitable is based on
| > the assumption that everyone lives in a city, and the technologies of
| > spying can be cheaply deployed.  A good deal of privacy can be
| > obtained by moving a small or large distance away.  Monitoring
| > technology is not cheap.  When it is cheap, the network links to
| > connect it all will still be expensive.  (etc.  The economics of a
| > surveillance state lead to something in the mix, people, cameras,
| > policemen to make arrests etc, being expensive.)

| As far as the unit cost of surveillance goes, it's cheaper every day.
| Hard disk (and other storage) space is way, way up per dollar, processing
| speeds and I/O are improving greatly per dollar, and the type of custom
| database software and O/S employed by the top surveillance pros is not
| at all analogous to the stuff most people use on Unix, DOS, or other
| common small computer systems.  I did some pioneer work in high-speed
| database work, and the software makes a BIG difference in unit cost of
| surveillance.

	Could you give me a cost estimate for keeping video of the
last 10 minutes of 250 million lives?  This is essentially one of
Brin's suggestions, and it strikes me as astoundingly pricey, even if
you just consider the cost of cameras, fiber, switches, and vcrs, and
ignore the problem of deciding what tape to keep.

	Some back of the envelope leads me to over a trillion,
figuring that a second of video takes 10kb, and disk costs about
$50/mb.  250m cameras at $40 each, fiber connections at $400 each,
etc.

Adam

-- 
"Pretty soon, you're talking about real money."