How Password Hashing Works

Password hashing is the process of running a user's password through a one-way function that produces a fixed-length output called a hash or digest. The original password cannot be recovered from the hash — that is the point.

When a user sets a password, you store the hash. When they log in, you hash their input again and compare it to the stored hash. If they match, the password is correct. The plaintext password is never stored.

Encryption is reversible. If you encrypt passwords and store them, you also have to store or manage the decryption key somewhere. An attacker who obtains your database may also find the key — or exploit the system that holds it. The result: all passwords exposed.

With hashing, there is no key to steal. Even if an attacker copies your entire password database, they cannot reverse the hashes directly. They must guess passwords one at a time and check if the hash matches — which is why the hashing algorithm's speed matters enormously.

MD5 and SHA-256 are general-purpose hash functions designed to be fast. That is the problem. Modern hardware can compute hundreds of millions or billions of SHA-256 hashes per second. If an attacker obtains your database of SHA-256 password hashes, they can systematically try common passwords at extreme speed.

Dedicated password hashers like bcrypt and Argon2 are intentionally slow. A single bcrypt operation at cost factor 12 takes roughly 250–400ms on a modern server. An attacker testing millions of guesses faces a genuine time barrier. Fast algorithms like SHA-256 offer no such barrier.

For a detailed comparison of MD5 and SHA-256 as general-purpose hashing tools, see MD5 vs SHA-256. Neither is appropriate for passwords.

A salt is a random string generated uniquely for each password before hashing. The salt is combined with the password before the hash function runs, and then stored alongside the hash.

Salting solves two problems. First, it ensures that two users with the same password produce different stored hashes — an attacker cannot tell they share a password. Second, it defeats precomputed lookup tables (rainbow tables), since any precomputed table would need to include every possible salt value.

Modern password hashing libraries (bcrypt, Argon2, PBKDF2) handle salt generation automatically. The salt is embedded in the output hash string. You do not need to manage it separately.

Key stretching is the technique of making a hash function intentionally slow by running it many times or requiring large amounts of memory. The parameter that controls this is called the work factor, cost factor, or iteration count depending on the algorithm.

The goal is to find a cost that is acceptable for legitimate logins (a few hundred milliseconds per user) but makes bulk offline attacks infeasible (millions of guesses would take years). Because hardware improves over time, you can increase the cost factor as needed — older hashes can be upgraded transparently when a user logs in.