How bcrypt Works – Java Password Hashing Security

Direct Answer

bcrypt is an adaptive password-hashing function built around the expensive Blowfish key setup. It generates a random salt, applies work controlled by a logarithmic cost, and stores version, cost, salt, and result in its standard 60-character string. Increasing the cost by one roughly doubles the work. bcrypt remains useful for existing systems, but it has a 72-byte password-input limit and OWASP prefers memory-hard Argon2id for new applications.

Code Example

@Bean
PasswordEncoder passwordEncoder() {
    int cost = measuredCostForThisService();
    return new BCryptPasswordEncoder(cost);
}

String stored = passwordEncoder.encode(candidate);
boolean accepted = passwordEncoder.matches(candidate, stored);

Choose the cost from a load test that includes peak concurrent logins. Record the chosen policy and use upgradeEncoding or an equivalent check to migrate older values after successful authentication.

Common Mistakes & Pitfalls

• Calling encode multiple times on the same password: Freshers sometimes do encoder.encode(encoder.encode(password)) thinking more hashing is more secure, but bcrypt output is 60 characters of mixed alphanumeric—hashing it again produces a hash of a hash, breaking the verification logic entirely.
• Using cost factor 4-6 for speed: Setting a low cost factor defeats the entire purpose of bcrypt. Cost 4 takes milliseconds and gives attackers the speed advantage you're trying to prevent. Start at 10 minimum, tune to 12-14 for production.
• Comparing hashes with equals: Never do storedHash.equals(encoder.encode(input))—this creates a new hash with a new salt, so it'll never match. Always use encoder.matches(), which extracts the salt from storedHash and recomputes correctly.
• Thinking bcrypt is encryption: Saying bcrypt encrypts passwords reveals a fundamental misunderstanding—encryption is reversible with a key, hashing is mathematically one-way.