How does challenge-response with symmetric cryptography authenticate a user without sending the password?

Both sides share a secret key. Server sends a random challenge; client returns HMAC(key, challenge); server verifies by computing the same HMAC.

The flow:

1. Setup: Alice and Bob share the same symmetric key (out-of-band — e.g. during account creation).
2. Login:
   • Alice → Bob: "I'm Alice"
   • Bob → Alice: random challenge (a fresh nonce, e.g. 128 bits)
   • Alice → Bob: response = HMAC(key, challenge)
   • Bob computes HMAC(key, challenge) himself and checks it matches the response.

Why it's better than password-in-the-clear:

• The secret key never leaves either machine.
• The challenge is random per session → an eavesdropper sees a different exchange every time and can't replay it.
• Even if Eve captures the response, it's only valid for that one challenge.

Drawback to remember: Bob also has a copy of the key. If Bob's server is compromised, the attacker can impersonate Alice — they have the same key Alice has. That's why public-key challenge-response is often preferred for high-stakes systems.