=
×

Julian Cope presents Head Heritage

Head To Head
Log In
Register
U-Know! Forum »
Calling Grufty Jim (& the rest of you 2)
Log In to post a reply
60 messages
Topic View: Flat | Threaded
grufty jim
grufty jim
1978 posts

Encryption as i see it
Nov 10, 2002, 16:12
Hey YAIP,

In your message you say you're "more technician than mathematician". Me, i'm the opposite. I can read and comprehend (after a few tries and a headache or two) the mathematics used by Einstein to describe Special Relativity, but a few minutes of staring at a car engine, say, and i'd probably start to hyperventilate (even though i can give you an accurate energy-systems analysis of that very engine! :-)

So i'm aware of the inherent strength of 2048 bit encryption. I understand how it works, and frankly believe that 128 bit is probably - for all practical purposes - currently uncrackable. Increasing key length is a way of increasing encryption complexity almost exponentially. So whilst a 56 bit key will only take 3 days using 100 networked computers; 109 bit will take a year and a half, using 10,000 computers and a professor of mathematics.

HOWEVER, and it's a *big* however; all of this is "today". Thanks to Moore's Law (which i notice has recently been extended yet again), but also to huge increases in the scope of distributed processing, the timescale for cracking any code is shortening every day. So in 18 months time (say), the same 109 bit code that it took 10,000 computers to crack in 18 months, will only require 5,000 much faster computers and will take half the time.

So although 128 bit encryption is secure today, i wouldn't bet on it still being secure in 10 years time. 1024/2048 bit encryption, however (and remember it's an exponential increase in key complexity - not a linear one) is probably secure for our likely lifetimes.

Only "probably" though, for one reason. Quantum computing. This is a subject i don't know much about, but the little i've read seems to suggest that quantum computing *may* allow incomprehensibly large increases in the ability of computers to perform mathematical analyses. We are still some way from quantum computing however (and frankly, i suspect we'll never see it outside laboratory prototypes for a bunch of reasons).

So until then, 2048 bit encryption is 100% reliable WHEN USED CORRECTLY AND SECURELY. There are dozens of possible human errors which could lead to even theoretical 4096 bit encryption being compromised - but the math doesn't lie, and is completely safe for now.
In reply to:
Re: Encryption (YerArseInParsley)		 Replies:
WHEN USED CORRECTLY AND SECURELY (YerArseInParsley)
Topic Outline:
U-Know! Forum Index