Paper 2
Paper 2
Paper 2
A possible implementation is suggested of J H Ellis's proposed method of encryption involving no sharing of secret
information (key lists, machine set-ups, pluggings etc) between sender and receiver.
1. In [1] J H Ellis describes a theoretical method of encryption which does not necessitate the sharing of secret
information between the sender and receiver. The following describes a possible implementation of this.
b. The sender has a message, consisting of numbers C1, C2, ... Cr with 0 < Ci < N
c. To decode, the receiver finds, by Euclid's Algorithm, numbers P', Q' satisfying
i. P P' = 1 (mod Q - 1)
Processes Involved
2. There is an algorithm, involving work of the order of log M, to test if M is prime, which usually works but can fail
to give an answer. Hence as the density of primes is (log M)-1, picking primes is a process of order (log M)k where
k is a small integer.
3. Also, computing CiN (mod N) is of order (log N)k' and the computation of DiP' and DiQ' even smaller; hence coding
and decoding is a process requiring work of order (log N)k where k will be about 2 or 3.
4. However, factorising N is a process requiring work of order N1/4 (log N)k, where k is a small integer (alternatively
computing C from CN (mod N) requires work of order N if the factorization of N is not known); so decoding for an
interceptor of the communication is a process of order about N1/4.
Reference [1] The possibility of Non-Secret digital encryption. J H Ellis, CESG Research Report, January 1970.
Note: There is no loss of security in transmitting C1 ... Cr all using the same N. Even if the enemy can guess a crib for
eg C1 ... Cr-1, this gives no information of use in decoding Dr etc. He could in any case provide himself with as many
pairs (Ci, Di) as he pleases, since the encryption process is known to him as well as to the transmitter!