Cryptographic failures
Cryptographic failures are failures related to cryptography, or the lack thereof, that often lead to the exposure of sensitive data or system compromise. Implementation mistakes often cause unexpected cryptographic errors.
First, let's begin by distinguishing between encoding, encryption, and hashing, in general programming terms.
- Encoding a value, converts the data to a different format for storage, transmission, compression, or decompression. Encoding helps transmit data in a channel, such as base 64 encoding over HTTP, but it doesn't provide security. It changes the format of a value but it doesn't protect the value from detection.
- Encryption is a reversible operation that translates text into what might seem like a random and meaningless cypher. To decrypt the value, an encryption key is needed.
- Hashing is a one-way operation of mapping input data into fixed-size values (value hash). There's no way to reverse hash a value.
Warning
Avoid writing your own cryptographic algorithms. Instead, use the strong cryptographic algorithms provided by .NET.
.NET framework provides you with all the tools you need out of the box. Including, encryption, hashing, and random number generation.
Your web applications deal with user accounts and data. But, how do you securely encrypt a secret, generate a password hash, or create a temporary password in your app? To answer this question, let's explore a couple of examples of what the System.Security.Cryptography namespace has to offer.
Encryption
To securely encrypt a value like a string or integer, symmetric or asymmetric encryption can be used. To encrypt data with a symmetric-key algorithm, we can use the Advanced Encryption Standard (AES). In the next example, a new instance of the Aes class is created and used to generate a new key and initialization vector (IV). The AES is used to encrypt any type of managed stream, then the stream is wrapped with CryptoStream.
Aes aes = Aes.Create();
CryptoStream cryptStream = new CryptoStream(fileStream,
aes.CreateEncryptor(aes.Key, aes.VI),
CryptoStreamMode. Write);
Hashing
Hashing is a one way operation. When using a hashing function to hash nonunique inputs such as passwords, use a salt value added to the original value before hashing.
public static byte[] HashPassword256(string password)
{
System.Security.Cryptography.SHA256 mySHA256 = System.Security.Cryptography.SHA256.Create();
var encoding = new System.Text.UnicodeEncoding();
return mySHA256.ComputeHash(encoding.GetBytes(password));
}
Random numbers
Temporary password or access codes are intended to be unique per user. The uniqueness can be achieved by introducing random character generation, and it's worth noting how the randomness is achieved. You might be familiar with the System.Random class. System.Random is a deterministic pseudo-random sequence generator. It’s predictable and seeded only from the system clock, which means it's guessable. In matter of fact, Microsoft Learn documentation explicitly states it shouldn't be used for generating passwords. To generate a cryptographically secure random number that's suitable for creating a random password, you should use a RandomNumberGenerator instance.
var randomNumberGenerator = System.Security.Cryptography.RandomNumberGenerator.Create();
By using RandomNumberGenerator, you can eliminate the chances of two or more users ending up with the same token or password, when they must be unique.
Code review notes
You don't want to unnecessarily store any sensitive data in your system. If you do, make sure you encrypt all data in transit and at rest with the use of HTTPS and TLS/SSL. After studying your teams codebase, you can now make the distinction between encryption, encoding, and hashing. No crypto keys were checked into the source code repository. You're planning to research security scanning tools to run as part of the CI/CD process to prevent sensitive data and secrets from getting into the repository.
Your app has a built-in mechanism for resetting user passwords. You noticed System.Security Cryptography.RandomNumberGenerator being used in favor of System.Random.