Re: .NET DES CFB - useless interface?
From: Peter Gutmann (pgut001_at_cs.auckland.ac.nz)
Date: 11/22/04
- Next message: Douglas A. Gwyn: "Re: shuffling algorithm"
- Previous message: Kristian Gjøsteen: "Re: AES - finding bits of incomplete key"
- In reply to: Dominic: ".NET DES CFB - useless interface?"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]
Date: 22 Nov 2004 09:04:58 GMT
dominicsmith501@hotmail.com (Dominic) writes:
>I always thought that the main point of CFB mode was that it allowed
>you to use a block cipher like a stream cipher.
>eg if you set the feedback size to 8 bits, you can put a byte in and
>get a byte out. etc
>The .NET framework implementation offers CFB mode for DES, and it even
>lets you set the feedback size.
>But unfortunately the TransformBlock method seems to insist that you
>pass in data which is an exact number of block lengths in size. And it
>also insists that the block length is 64 bits (ie the block size for
>DES).
>This basically makes it impossible to use CFB mode to stream data in
>any smaller chunks than you could do with CBC anyway.
That problem dates back to CryptoAPI 1.0, circa 1996, and yes, it does make
the stream modes completely useless. I don't know about .NET, but in
CryptoAPI the feedback also defaults to CFB-8 rather than CFB-64.
Peter.
- Next message: Douglas A. Gwyn: "Re: shuffling algorithm"
- Previous message: Kristian Gjøsteen: "Re: AES - finding bits of incomplete key"
- In reply to: Dominic: ".NET DES CFB - useless interface?"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]
Relevant Pages
|