Logo Questions Linux Laravel Mysql Ubuntu Git Menu
 

What is the easiest way to encrypt a password when I save it to the registry?

People also ask

Which encryption is best for passwords?

Google recommends using stronger hashing algorithms such as SHA-256 and SHA-3. Other options commonly used in practice are bcrypt , scrypt , among many others that you can find in this list of cryptographic algorithms.

What command is used to encrypt all passwords?

To help for security and reduce password vulnerability, the command _service password-encryption encrypts passwords so that they cannot be displayed in plain text in a printed version or backup copy of the configuration file.


You don't decrypt authentication passwords!

Hash them using something like the SHA256 provider and when you have to challenge, hash the input from the user and see if the two hashes match.

byte[] data = System.Text.Encoding.ASCII.GetBytes(inputString);
data = new System.Security.Cryptography.SHA256Managed().ComputeHash(data);
String hash = System.Text.Encoding.ASCII.GetString(data);

Leaving passwords reversible is a really horrible model.

Edit2: I thought we were just talking about front-line authentication. Sure there are cases where you want to encrypt passwords for other things that need to be reversible but there should be a 1-way lock on top of it all (with a very few exceptions).

I've upgraded the hashing algorithm but for the best possible strength you want to keep a private salt and add that to your input before hashing it. You would do this again when you compare. This adds another layer making it even harder for somebody to reverse.


Please also consider "salting" your hash (not a culinary concept!). Basically, that means appending some random text to the password before you hash it.

"The salt value helps to slow an attacker perform a dictionary attack should your credential store be compromised, giving you additional time to detect and react to the compromise."

To store password hashes:

a) Generate a random salt value:

byte[] salt = new byte[32];
System.Security.Cryptography.RNGCryptoServiceProvider.Create().GetBytes(salt);

b) Append the salt to the password.

// Convert the plain string pwd into bytes
byte[] plainTextBytes = System.Text UnicodeEncoding.Unicode.GetBytes(plainText);
// Append salt to pwd before hashing
byte[] combinedBytes = new byte[plainTextBytes.Length + salt.Length];
System.Buffer.BlockCopy(plainTextBytes, 0, combinedBytes, 0, plainTextBytes.Length);
System.Buffer.BlockCopy(salt, 0, combinedBytes, plainTextBytes.Length, salt.Length);

c) Hash the combined password & salt:

// Create hash for the pwd+salt
System.Security.Cryptography.HashAlgorithm hashAlgo = new System.Security.Cryptography.SHA256Managed();
byte[] hash = hashAlgo.ComputeHash(combinedBytes);

d) Append the salt to the resultant hash.

// Append the salt to the hash
byte[] hashPlusSalt = new byte[hash.Length + salt.Length];
System.Buffer.BlockCopy(hash, 0, hashPlusSalt, 0, hash.Length);
System.Buffer.BlockCopy(salt, 0, hashPlusSalt, hash.Length, salt.Length);

e) Store the result in your user store database.

This approach means you don't need to store the salt separately and then recompute the hash using the salt value and the plaintext password value obtained from the user.

Edit: As raw computing power becomes cheaper and faster, the value of hashing -- or salting hashes -- has declined. Jeff Atwood has an excellent 2012 update too lengthy to repeat in its entirety here which states:

This (using salted hashes) will provide the illusion of security more than any actual security. Since you need both the salt and the choice of hash algorithm to generate the hash, and to check the hash, it's unlikely an attacker would have one but not the other. If you've been compromised to the point that an attacker has your password database, it's reasonable to assume they either have or can get your secret, hidden salt.

The first rule of security is to always assume and plan for the worst. Should you use a salt, ideally a random salt for each user? Sure, it's definitely a good practice, and at the very least it lets you disambiguate two users who have the same password. But these days, salts alone can no longer save you from a person willing to spend a few thousand dollars on video card hardware, and if you think they can, you're in trouble.


Tom Scott got it right in his coverage of how (not) to store passwords, on Computerphile.

https://www.youtube.com/watch?v=8ZtInClXe1Q

  1. If you can at all avoid it, do not try to store passwords yourself. Use a separate, pre-established, trustworthy user authentication platform (e.g.: OAuth providers, you company's Active Directory domain, etc.) instead.

  2. If you must store passwords, don't follow any of the guidance here. At least, not without also consulting more recent and reputable publications applicable to your language of choice.

There's certainly a lot of smart people here, and probably even some good guidance given. But the odds are strong that, by the time you read this, all of the answers here (including this one) will already be outdated.


The right way to store passwords changes over time.

Probably more frequently than some people change their underwear.


All that said, here's some general guidance that will hopefully remain useful for awhile.

  1. Don't encrypt passwords. Any storage method that allows recovery of the stored data is inherently insecure for the purpose of holding passwords - all forms of encryption included.
  2. Process the passwords exactly as entered by the user during the creation process. Anything you do to the password before sending it to the cryptography module will probably just weaken it. Doing any of the following also just adds complexity to the password storage & verification process, which could cause other problems (perhaps even introduce vulnerabilities) down the road.

    • Don't convert to all-uppercase/all-lowercase.
    • Don't remove whitespace.
    • Don't strip unacceptable characters or strings.
    • Don't change the text encoding.
    • Don't do any character or string substitutions.
    • Don't truncate passwords of any length.
  3. Reject creation of any passwords that can't be stored without modification. Reinforcing the above. If there's some reason your password storage mechanism can't appropriately handle certain characters, whitespaces, strings, or password lengths, then return an error and let the user know about the system's limitations so they can retry with a password that fits within them. For a better user experience, make a list of those limitations accessible to the user up-front. Don't even worry about, let alone bother, hiding the list from attackers - they'll figure it out easily enough on their own anyway.

  4. Use a long, random, and unique salt for each account. No two accounts' passwords should ever look the same in storage, even if the passwords are actually identical.
  5. Use slow and cryptographically strong hashing algorithms that are designed for use with passwords. MD5 is certainly out. SHA-1/SHA-2 are no-go. But I'm not going to tell you what you should use here either. (See the first #2 bullet in this post.)
  6. Iterate as much as you can tolerate. While your system might have better things to do with its processor cycles than hash passwords all day, the people who will be cracking your passwords have systems that don't. Make it as hard on them as you can, without quite making it "too hard" on you.

Most importantly...

Don't just listen to anyone here.

Go look up a reputable and very recent publication on the proper methods of password storage for your language of choice. Actually, you should find multiple recent publications from multiple separate sources that are in agreement before you settle on one method.

It's extremely possible that everything that everyone here (myself included) has said has already been superseded by better technologies or rendered insecure by newly developed attack methods. Go find something that's more probably not.


This is what you would like to do:

OurKey.SetValue("Password", StringEncryptor.EncryptString(textBoxPassword.Text));
OurKey.GetValue("Password", StringEncryptor.DecryptString(textBoxPassword.Text));

You can do that with this the following classes. This class is a generic class is the client endpoint. It enables IOC of various encryption algorithms using Ninject.

public class StringEncryptor
{
    private static IKernel _kernel;

    static StringEncryptor()
    {
        _kernel = new StandardKernel(new EncryptionModule());
    }

    public static string EncryptString(string plainText)
    {
        return _kernel.Get<IStringEncryptor>().EncryptString(plainText);
    }

    public static string DecryptString(string encryptedText)
    {
        return _kernel.Get<IStringEncryptor>().DecryptString(encryptedText);
    }
}

This next class is the ninject class that allows you to inject the various algorithms:

public class EncryptionModule : StandardModule
{
    public override void Load()
    {
        Bind<IStringEncryptor>().To<TripleDESStringEncryptor>();
    }
}

This is the interface that any algorithm needs to implement to encrypt/decrypt strings:

public interface IStringEncryptor
{
    string EncryptString(string plainText);
    string DecryptString(string encryptedText);
}

This is a implementation using the TripleDES algorithm:

public class TripleDESStringEncryptor : IStringEncryptor
{
    private byte[] _key;
    private byte[] _iv;
    private TripleDESCryptoServiceProvider _provider;

    public TripleDESStringEncryptor()
    {
        _key = System.Text.ASCIIEncoding.ASCII.GetBytes("GSYAHAGCBDUUADIADKOPAAAW");
        _iv = System.Text.ASCIIEncoding.ASCII.GetBytes("USAZBGAW");
        _provider = new TripleDESCryptoServiceProvider();
    }

    #region IStringEncryptor Members

    public string EncryptString(string plainText)
    {
        return Transform(plainText, _provider.CreateEncryptor(_key, _iv));
    }

    public string DecryptString(string encryptedText)
    {
        return Transform(encryptedText, _provider.CreateDecryptor(_key, _iv));
    }

    #endregion

    private string Transform(string text, ICryptoTransform transform)
    {
        if (text == null)
        {
            return null;
        }
        using (MemoryStream stream = new MemoryStream())
        {
            using (CryptoStream cryptoStream = new CryptoStream(stream, transform, CryptoStreamMode.Write))
            {
                byte[] input = Encoding.Default.GetBytes(text);
                cryptoStream.Write(input, 0, input.Length);
                cryptoStream.FlushFinalBlock();

                return Encoding.Default.GetString(stream.ToArray());
            }
        }
    }
}

You can watch my video and download the code for this at : http://www.wrightin.gs/2008/11/how-to-encryptdecrypt-sensitive-column-contents-in-nhibernateactive-record-video.html