From: Christian Weisgerber (naddy@mips.inka.de)
Date: Mon Jun 23 2003 - 10:35:20 MDT
Brett Paatsch <paatschb@optusnet.com.au> wrote:
> I'm no encryption expert but since reading Simon Singh's
> The Code Book back in 1999 I've had the impression that
> there is a race on between the developers of quantum
> computers that could crack existing codes and the development
> of quantum encryption which could produce uncrackable
> codes even given quantum computers. I'm not sure what the
> state of the art is - the following article seems to gives some hints.
As Eugen Leitl likes to quip, it's math versus physics and he prefers
to trust the math.
Nowadays, transmissions are routinely secured by encryption which
is mathematical. To send a message, you send a bunch of data. It
can pass through many different networks, over different media,
etc. As long as the data eventually arrives, encryption is applicable.
It's quite possible that practical quantum computers (which we don't
have and won't have in the immediate future) might greatly reduce
the effort required to break current encryption schemes. Which
doesn't necessarily mean that it will be trivial or even practical.
Anyway, people don't worry much about this, because it stands to
reason that by the time practical quantum computers are available,
there will also be new encryption schemes that take advantage of
the properties of quantum computing and will be just as unbreakable
with quantum computers as current schemes are with current hardware.
What nowadays keeps popping up as "quantum encryption" (e.g. in the
article you quoted) is something else entirely. It refers to
securing a transmission against eavesdropping by making use of
quantum properties. In practice, it relies on the ability to
transmit single photons from sender to recipient. It _only_ secures
a physical transmission line by relying on physical properties.
Since single photons tend to get lost but sending a group would
negate the whole concept, this is rather tricky to implement.
To safeguard the message, sender and recipient must be connected
by a single physical link, e.g. line-of-sight laser or a fiberoptic
cable. You cannot forward it from optical to electric and through
different networks.
> If it was possible to put quantum encryption affordably on a
> laptop yet effectively everyone would have it would that be a
> net good?
"Quantum encryption" is simply not applicable to this. If Brett
and I had such a module in our laptops, we would still need to run
a fiber line between our homes in Germany and Australia to securely
talk to each other. (Actually, the current range limit for laboratory
stage "quantum encryption" is 100km.)
Our modern global communications infrastructure is the antithesis
to this. Say, I want to send an IP packet from my computer to
yours. From computer to switch (ethernet), from switch to dial-up
router (ethernet), from router to NT (ISN S0 bus), from NT to CO
(ISDN U interface). By the time the packet is out of my home it's
already on its fourth physical link. Any long distance
Internet/telephony/etc data exchange will cross dozens of physical
links. No problem for encryption, which can be handled end-to-end.
> Privacy for all. Or privacy for governments only? Thoughts?
That's worth some thoughts, but "quantum encryption" does not impinge
on this.
-- Christian "naddy" Weisgerber naddy@mips.inka.de
This archive was generated by hypermail 2.1.5 : Mon Jun 23 2003 - 11:44:04 MDT