Enigma was the most famous cipher ever broken, and it was broken while it was still in service. This page is about why — because the reasons are still the reasons we build security the way we do.
Why "it's old" is not the reason
It is tempting to say a 1920s machine is insecure because it is old, as though age were the defect. The useful question is what a cipher has to guarantee, and Enigma fails three requirements that any modern one takes for granted.
1. Keys were reused
Operators were issued a key sheet a day (a month at sea) and set the same rotors, rings, positions and plugboard pairs for every message in it. One key protected thousands of letters — which is exactly the condition that makes statistical analysis work, because the language underneath starts to show through the cipher. Modern cryptography assumes the key is used for a small, fixed amount of data; when that amount grows, the cipher stops being safe.
2. The plaintext was guessable
Military radio traffic had set phrases in predictable places — a routine weather report, an opening formula — so an analyst could guesspart of the message, put the guess and the ciphertext side by side, and learn something about the key. This is crib dragging, and it is devastating against Enigma because the plugboard pairs can be recovered one at a time from a good crib. It also made the cipher malleable: an attacker could insert a chosen phrase, light the matching lamp and change the meaning of a message without breaking it. A modern cipher is authenticated precisely so that a modified message is detectable.
3. The search space was small
Once the plugboard was known, the remaining key was the rotor order (336 choices with no repeated rotor), the starting positions (26³ = 17,576) and the reflector (3) — about 18 million combinations altogether. The British bombe, designed by Alan Turing with a vital correction from Welchman, was built to search that spacemechanically because checking it by hand was hopeless.
Today that same search takes seconds on a laptop, which is the point: a cipher's security has to hold against an adversary with a modern computer and unlimited time on the ciphertext. Cryptography that depended on the opponent's equipment being slow was never going to survive.
What the war taught cryptanalysts
- Key management is most of the work. The maths was sound; the practice — one key per day, written down, carried by hand — was what failed. This is still true, and it is why modern systems spend more code on key exchange than on encryption.
- Attack the protocol, not just the cipher. Crib dragging exploited a weakness in how the machine was used, not in the wiring. Padding-oracle and related-key attacks are the same move a century later.
- Public scrutiny makes stronger ciphers. Encryption whose security depends on keeping the design secret has to be trusted; ciphers that are openly specified and analysed are the ones that get better. We no longer accept "trust us, it's secure".
- Shannon (1945) defined the target. His perfect secrecy — the ciphertext alone tells an attacker nothing — is achievable only with a one-time pad. Every practical cipher since is a deliberate compromise: slightly less than perfect, but possible to use.
What replaced it
- Fast symmetric ciphers. DES (1977) and then AES (2001) do with software what Enigma did with rotors, and are designed to withstand public attack. AES-256 or ChaCha20-Poly1305 are what a modern browser negotiates.
- Public-key cryptography. Diffie-Hellman (1976) and RSA (1977) solved Enigma's other impossible problem: agreeing a secret key over a channel you cannot trust. You can now set up shared secrets with someone you have never met.
- Authenticated encryption. Confidentiality alone is not enough — a message also has to be provably unaltered. Modern protocols seal ciphertext with a key (AEAD) so tampering fails loudly instead of quietly, which is the direct answer to the lampboard tampering of 1943.
- Real randomness and real protocols. Keys come from cryptographically secure random sources, key exchange happens forward secrecy (past sessions cannot be recovered from a captured one), and transport security is TLS rather than a cipher you chose yourself.
The short version
- Use an open, reviewed algorithm — never one you designed yourself.
- Never reuse a key; rotate it, and keep it secret for its whole life.
- Choose key length for the security you need, not for the algorithm's reputation.
- Authenticate as well as encrypt, and check who you are talking to.
- Take randomness from a secure source, not from a clock or a counter.
None of this makes Enigma worthless — it makes it a machine worth understanding. You can still operate it on thesimulator, with the wirings the Wehrmacht and Kriegsmarine actually used, and watch for yourself how little the machine knows about the person on the other end of the wire.