What Are the Optimal Parameters for Scrypt Work Factors in Password Storage?

What are the optimal scrypt work factors for secure password storage?

The scrypt algorithm is designed to provide a secure way to hash passwords by making the hashing process resource-intensive to deter brute-force attacks. It uses three parameters: N (CPU cost), r (memory cost), and p (parallelization cost), each of which plays a critical role in determining how resistant a password hash is to cracking attempts.

import org.bouncycastle.crypto.generators.SCrypt;

public class PasswordHasher {
    public static byte[] hashPassword(String password, byte[] salt) {
        int N = 16384; // CPU cost
        int r = 8;     // memory cost
        int p = 1;     // parallelization cost
        int dkLen = 32; // output length in bytes
        return SCrypt.generate(password.getBytes(), salt, N, r, p, dkLen);
    }
}

Causes

• The choice of parameters directly affects the security and performance of password hashing.
• Insufficient parameters can lead to vulnerabilities against modern brute-force attacks.

Solutions

• **N (CPU cost)**: This parameter can be set to 16384 or higher. As a general guideline, use powers of 2 for N, such as 8192, 16384, 32768, etc. Higher values increase security but require more processing power.
• **r (memory cost)**: Commonly set between 6 and 8. This value can dictate how much memory is used during the hashing process. Increasing it also helps protect against hardware attacks.
• **p (parallelization cost)**: A value of 1 is often sufficient unless you are dealing with a scenario where you have multiple threads needing to generate hashes simultaneously. If that’s the case, consider using a higher value.

Mistake: Using too low of a value for N, r, or p, making hashes vulnerable to attacks.

Solution: Always test against known attacks and validate against standards; start from recommended values like N=16384, r=8, p=1.

Mistake: Inadequate understanding of what N, r, and p mean and their implications.

Solution: Research relevant literature or use community recommendations such as those from OWASP to better understand these parameters.