The GCHQ Trio — Ellis, Cocks, and Williamson
They invented RSA and Diffie-Hellman first. No one knew for 27 years.
Invention Biography
The GCHQ Trio — Ellis, Cocks, and Williamson
James Ellis’s classified report on the possibility of non-secret encryption is dated January 1970. Clifford Cocks proposed an RSA-related construction in 1973, and Malcolm Williamson developed finite-field key agreement in 1974. Their work remained classified while public academic research independently developed public-key cryptography. Public disclosure of the GCHQ work came in 1997.
Why This Person Matters
The episode distinguishes private discovery from publication and public adoption. Public-key methods became central to internet authentication and key establishment, but a mathematical proposal is not the same as a deployed, authenticated, securely encoded protocol. Counterfactual claims that earlier publication would have advanced the entire field by an exact number of years cannot be established.
Ellis’s January 1970 report, The Possibility of Secure Non-Secret Digital Encryption, argued that the idea was possible without supplying a practical construction. This date is the report’s date, distinct from accounts of his earlier thought process.
Cocks’s 1973 work used modular exponentiation with a composite modulus, closely related to the later RSA construction. Knowledge of the factors enables the needed private computation. It is not established that every possible attack on RSA inversion must factor the modulus, nor that Cocks’s formulation and every later RSA protocol were identical in every detail.
GCHQ declassified the Ellis and Cocks papers in 1997, after RSA Security had already licensed RSA as a patented system and built a company around it. Cocks received a Companion of the Order of St Michael and St George (CMG) in 2008. Ellis died in November 1997, just weeks before the public disclosure — he never saw the world learn his contribution. The academic cryptography community's response was universal acknowledgment: the published RSA paper would have acknowledged Ellis and Cocks had their work been known.
| James Ellis | 1924–1997 — proposed non-secret encryption concept, report dated January 1970 |
| Clifford Cocks | b. 1950 — implemented RSA equivalent, 1973 |
| Malcolm Williamson | b. 1950 — implemented Diffie-Hellman equivalent, 1974 |
| Organization | GCHQ (Government Communications Headquarters), Cheltenham |
| Classification | Top secret until 1997 |
| Public credit | Diffie & Hellman (1976), Rivest, Shamir & Adleman (1977) |
| Recognition | Clifford Cocks: CMG 2008 |