[Enigmail] First signed message -- Is tinyurl.com safe and reliable?
Robert J. Hansen
rjh at sixdemonbag.org
Mon Aug 25 04:11:41 PDT 2008
Phil Stracchino wrote:
> Core memory and punch-tape readers, whee!
Off topic, but kind of funny --
Years ago I was taking the graduate series in computer security. One
question on the final exam was something as follows:
"You are a virus author who has managed to get your code running on
untrusted hardware within your target's perimeter. Since your host
machine is untrusted, the RAM is scanned by an uninfected process on a
different machine every thirty seconds to ensure its integrity. The
scan requires two seconds. Your computation task will require five
minutes of computing. How will your virus avoid ten independent virus
scans, potentially with ten different detection algorithms?"
=====
I read the question, then walked up to the proctor's desk, where Dr.
Jones was reading the latest _Communications of the ACM_. "Can we
assume we have total control over the software for thirty seconds?" I
asked. Dr. Jones nodded.
"Do you mind if I hypothesize some exotic networking?"
Dr. Jones shook his head no. As long as it was firmly rooted in modern
technological capability, he didn't care if my attack used fibre
channel, ATM, token ring or anything else.
So I went back and drafted my answer.
=====
"My process will do a partial computation. It will then beam itself out
of the system, encoding both itself and its data in a laser pulse aimed
at the Moon. The last thing it does on its way out is clean up after
itself. Once it hits the Moon, it hits the reflector array left behind
by the Apollo astronauts, and bounces back to Earth. The 2.6-second
round trip time is enough to allow the virus detector to run. When the
pulse returns to Earth, it's picked up by the network and re-infects the
system all over again, restoring it to the point in computation it was
at before it left."
I received full credit for the answer. Dr. Jones presented it in class
as an example of an "unorthodox" and "too complicated" solution -- he
said it would have been far more effective just to bounce it around the
internet -- but as one that was technically correct.
The class, of course, was horrified. One person turned to me, scowled,
and loudly said "EINSTEINIAN SPACETIME IS NOT A STORAGE MEDIUM!" (I
couldn't help but howl at that. Well, um, _usually_ it's not.)
To which Jones just muttered something about he didn't understand why
people thought this was either (a) creative or (b) offensive to the
sensibilities, since it was just a really long-path delay-line memory
cell, and we've had those since the 50s...
Everything that's old is new again. :)
http://en.wikipedia.org/wiki/Delay_line_memory
More information about the Enigmail
mailing list