- From: Joseph Ashwood <ashwood@msn.com>
- Date: Thu, 20 Jun 2002 16:14:36 -0700
- To: <xml-encryption@w3.org>
----- Original Message -----
From: "Tom Gindin" <tgindin@us.ibm.com>
> The major reason IMHO why anyone would want to use some of these
> algorithms which are 64 bits or less would be that no stronger algorithm
is
> available in both the encryptor's environment and the decryptor's.
Since 128-bit Rijndael/AES is required, that situation can never occur,
there will always be a 128-bit algorithm available.
> The
> legacy that matters in this case consists of legacy crypto libraries since
> obviously there's no legacy of XML-encrypted documents.
With Rijndael as a requirement, the presence of a legacy crypto library is
unlikely. With the additional observation that libraries like OpenSSL and
Crypto++ are available free it is even less likely that legacy crypto
libraries will be used.
>DES is also faster
> than 3DES, of course, but that's not a very strong reason.
And Rijndael is faster than either, more secure, and one of the required
algorithms.
> My own point had to do with PBE because of its key exchange and
> storage characteristics. I don't think we should try to add every PBE
> variant in PKCS#5 and PKCS#12, let alone all others which have ever been
> defined.
PBEs might be useful, but I think they should focus on the algorithms that
are already significant options, like Rijndael/AES, 3DES, etc. This would
reduce the code necessary for an implementation, reduce the executable size,
and shrink the specification, and won't compromise security or the speed.
Joe
Received on Thursday, 20 June 2002 19:15:35 UTC