bcrypt

What it means

bcrypt is a password hashing function designed in 1999 by Niels Provos and David Mazières, built on the Blowfish cipher. It is deliberately slow, and you can make it slower as computers get faster by raising a setting called the cost factor. It generates a random salt for each password and stores it inside the output.

The result is a single text string, commonly starting $2b$, followed by the cost, the salt and the hash. Verifying a password means reading the cost and salt from that string, hashing the attempt the same way, and comparing. Each increase of one in the cost doubles the work.

Why it matters

A fast hash lets an attacker with a stolen database try enormous numbers of guesses. bcrypt slows each guess, and the built-in salt stops one guess being checked against every account at once or against a precomputed table. A weak password still falls quickly, but a strong one gets a real defence.

It has been used widely for decades and is supported in nearly every language, so it is a safe, familiar choice. It is not memory-hard, though, so specialised hardware can speed attacks more than for Argon2, which is why current guidance prefers the newer design where available.

How it shows up

You will see it in the user table as 60-character strings. OWASP suggests a work factor of at least 10 and more if your servers allow, and says to raise it over time. Check the delay at login: it should be noticeable to an attacker but not annoying to a user, typically a fraction of a second.

The known trap is the 72-byte input limit. Longer passwords are cut off or, in some libraries, rejected, so two long passwords sharing their first 72 bytes can look identical. Other mistakes: leaving the cost at an old default, hashing the password with a fast hash first without care, and failing to upgrade old hashes at the next sign-in. See also key derivation functions.