[.net core學習] .net core中的Rijndael取代方法

SZW發表於2017-02-28

The difference (in .NET) between Rijndael and AES is that Rijndael allows the block size to change, but AES does not. Since RijndaelManaged's default block size is the same as the AES block size (128 bit / 16 byte) you are, in fact, using AES.

Instead of instantiating the implementation type by name, just use the factory (Aes.Create()). That works in both .NET Core and .NET Framework.

Other things worth mentioning:

  • All SymmetricAlgorithm instances are IDisposable, you should use them in a usingstatement.
  • All ICryptoTransform instances (such as your incorrectly named desEncryptor) are IDisposable, you should use them in a using statement.
  • ISO10126 padding is not available in .NET Core 1.0. If you need to be compatible with existing streams you can apply the padding yourself and specify PaddingMode.None. Otherwise, PKCS7 is more standard.
  • Your AES key isn't very random, since it comes from an ASCII string (lots of values won't be valid).
    • Base64 at least has full value range
    • PBKDF2 (Password-Based Key Derivation Function 2) via the Rfc2898DeriveBytes class allows for shared-string-secret in, predictable noise out.
    • KeyAgreement is in general better, but neither ECDH nor classic DH are available in .NET Core 1.0.
  • Usually the encryptor should let a random IV be calculated (call aes.GenerateIV() if using the same object for multiple operations) and present it with the ciphertext. So encrypt takes a key and plaintext and produces a ciphertext and IV. Decrypt takes (key, IV, ciphertext) and produces plaintext.

 

用法: SymmetricAlgorithm des = Aes.Create();

 

參考:http://stackoverflow.com/questions/38333722/how-to-use-rijndael-encryption-with-a-net-core-class-library-not-net-framewo

 

相關文章