James H. Ellis - Invention of Non-secret Encryption

Invention of Non-secret Encryption

Ellis said that the idea first occurred to him after reading a paper from World War II by someone at Bell Labs describing a way to protect voice communications by the receiver adding (and then later subtracting) random noise (probably the 1945 paper co-authored by Claude Shannon). He realised that 'noise' could be applied mathematically but was unable to devise a way to implement the idea.

Shortly after joining GCHQ in September 1973, after studying Mathematics at Cambridge University, Clifford Cocks was told of Ellis' proof and that no one had been able to figure out a way to implement it. He went home, thought about it, and returned with the basic idea for what has become known as the RSA asymmetric key encryption algorithm. Because any new and potentially beneficial/harmful technique developed by GCHQ is by definition classified information, the discovery was kept secret.

Not long thereafter, Cocks' friend and fellow mathematician, Malcolm Williamson, now also working at GCHQ, after being told of Cocks' and Ellis' work, thought about the problem of key distribution and developed what has since become known as Diffie–Hellman key exchange. Again, this discovery was classified information and it was therefore kept secret.

When, a few years later, Diffie and Hellman published their 1976 paper, and shortly after that Rivest, Shamir, and Adleman announced their algorithm, Cocks, Ellis, and Williamson suggested that GCHQ announce that they had previously developed both. GCHQ decided against publication at the time.

At this point, only GCHQ and the National Security Agency (NSA) in the USA knew about the work of Ellis, Cocks and Williamson. Whitfield Diffie heard a rumour, probably from the NSA, and travelled to see James Ellis. The two men talked about a range of subjects until, at the end, Diffie asked Ellis "Tell me how you invented public-key cryptography". After a long pause, Ellis replied "Well, I don't know how much I should say. Let me just say that you people made much more of it than we did."

On 18 December 1997, Clifford Cocks delivered a public talk which contained a brief history of GCHQ's contribution so that Ellis, Cocks and Williamson received some acknowledgment after nearly three decades of secrecy. James Ellis died on 25 November 1997, a month before the public announcement was made.