Prev: Modes of operation
Next: is the PGP signature format the most widely supported one?
From: rossum on 10 Mar 2010 09:06
On Wed, 10 Mar 2010 02:18:20 -0800 (PST), adacrypt
<austin.obyrne(a)hotmail.com> wrote:

>
>Alice writes an encryption program first of all. She then writes a
>corresponding decryption program that checks her previous encryption
>work. She tweaks these to perfection and then calls this joint
>combination of encryption and decryption programs her 'server'
Fine. I can follow that.

>that she now sends to Bob.
How? Alice needs a secure channel to Bob along which to send her
server. If the channel is not secure then we can assume that Eve, the
attacker, also has a copy of the server.

>In future she will communicate with Bob by
>sending him markup code as cipher text that will index her server
>(functioning now as an interpreter program) and Bob's computer then
>becomes his browser that displays the message that Alice wants him to
>know. This is a figurative model of what exists already as up-and-
>running working ciphers in my computer.
>
>The arrays of Alice's base interpreter program can be made as large as
>she likes so that the number of workable permutations of the order of
>the elements is literally out of this world in magnitude. The
>particular permutation that she initially sends to any Bob however is
>unique to that Bob and is simply one only element of this vast set.
>She can periodically refresh her 'interpreter' to this Bob by
>occasionally sending him external scrambling and slicing parameters
>that he must immediately apply to her particular server that is in his
>sole custody. In all her dealings with this Bob it is this same base
>that is being scrambled and sliced all the time. A different Bob
>would have had a different base interpreter sent to him initially so
>that the parameters of different messages (cross-channels) are useless
>if the cipher text of different messages are illegally intercepted by
>an adversary.
How does this differ from an insecure Two-Time-Pad? How does this
give us any advantage over a Maurer-style stream cypher that uses a
very large public database and hence has no need to maintain two
separate databases?

rossum

From: rossum on 10 Mar 2010 11:33
On Wed, 10 Mar 2010 07:28:17 -0800 (PST), adacrypt
<austin.obyrne(a)hotmail.com> wrote:

>This cryptography is so secure that even if the cost of the initial
>secure delivery was very great it would still be attractive
This cryptography is provably less secure than a One Time Pad and has
the same distribution problem. What advantage does it have over the
One Time Pad?

You also did not answer my point about a Maurer-style stream cypher
which gets round the distribution problem albeit at some cost in
security.

rossum

 | 
Pages: 1
Prev: Modes of operation
Next: is the PGP signature format the most widely supported one?