| From: | The Powerhouse <P.C.Steele@*********.AC.UK> |
|---|---|
| Subject: | Optical comms and encryption |
| Date: | Thu, 4 Aug 1994 17:14:58 +0100 |
Optical comms.
The idea of quantum encryption will not work unless you use only 1 cable and
have some device which can distinguish between photons at the receiving end.
NB such a device probably exists in places like CERN but the real world uses
PIN photodiodes to detect photons which don't necessarily detect 'every' photon.
They do get the majority though.
Other reasons why it would probably never work, even with one cable.
Attenuation : All optical fibres suffer from it in some form or another, light
is lost to the surrounding casing which can eventually be lost into the
cladding, this effect is more pronounced on corners.
Why it wouldn't work on any large network. As someone mentioned, repeaters
are used, virtual circuits are used, the whole telephony switching thing is
not constant. A signal may be represented in photons for the majority of it's
journey but as soon as it hits some relay station you'll lose whatever
encryption you were relying on as the message is encoded into a set of data
stored on a high speed computer ready to be retransmitted/rerouted.
I could go on and on but needless to say that the idea with todays technology
is not really going to be possible over any distance. It might be possible for
small distances <less than 100m> but then you lose the advantage of the
communication a) becasue the distance is negligble and b) as any site which can
afford this level of security will have more reliable site security and hence
doesn't need to worry about this level of communications security.
One time pads.
These are the only known form that cannot be cracked. Basically because you
use one unique set of data per message and there is no mathmatical formula
behind the code. One time pads only become vulnerable when they are used more
than once, ie they are no longer one time pads.
Other encryption.
As to the key being longer than the message. This certainly sounds feasible
and would fit in with my very basic knowledge of encryption, but as somone
said, this is useless for real time comms <ie telephony> which is where the
real push for secure encryption is.
Other types of encryption such as The Skipjack algorithm <not that I've seen
it :) > and RSA use a mathmatical formula which is impossible to find the
inverse of. The RSA one is well documented and involves taking the MOD of
various numbers and requires the sending party and the receiving party to
have a public and a private key. The Skipjack algorithm I hear also works in
the same way but has the extra facility of a Law Enforcement Access Field
<LEAF>, though even this can be subverted <although in a rather cumbersome
way> :)
Phill.
--
Phillip Steele - Email address P.C.Steele@***.ac.uk | Fighting against
Department Of Electrical & Electronic Engineering | Political Correctness !
University Of Newcastle Upon Tyne, England |
Land of the mad Geordies | The Powerhouse
