- From: Blair Dillaway <blaird@microsoft.com>
- Date: Thu, 17 May 2001 13:38:36 -0700
- To: "Al Sutton" <al@alsutton.com>, "Joseph M. Reagle Jr." <reagle@w3.org>
- Cc: <xml-encryption@w3.org>, "Philippe Le Hegaret" <plh@w3.org>
The draft specification provides info on this, though it may not be very clearly worded in the current draft. The EncryptedData elements contain the octet sequence corresponding to the encrypted, serialized representation of one or more child nodes of the element <1>. That is all the recipient knows. In the below example, the first encrypted data holds E("<a/>"), where E is the encryption transform. When you decrypt, you get back the XML source of "<a/>". You then need to parse this source to get back the child element <a/>. But, the first EncryptedData could just as easily contain E("here is some element text<a/><c/>"). The recipient won't know until they decrypt. Depending on the schema for the element <1>, the presence of 2 EncryptedData child element may, or may not, give you useful knowledge about what has been encrypted. Possible known plain-text attacks on the EncyrptedData is something the WG has agreed will be discussed in the specification. Hope this explaination helps. -----Original Message----- From: Al Sutton [mailto:al@alsutton.com] Sent: Thursday, May 17, 2001 11:24 AM To: Blair Dillaway; Joseph M. Reagle Jr. Cc: xml-encryption@w3.org; Philippe Le Hegaret Subject: Re: Comments on the 6 Apr Draft Hi, I've just read the last message in this thread and I have a couple of questions. If <1> <a/> <b/> </1> goes to... <1> <EncryptedData/> <EncryptedData/> </1> Then how can a and b be reconstructed as a and b during the decryption process?, and if there is some method then shouldn't the encrypted version of the children not give any clue as to the number of children that are encyrpted in order to limit the information given to a potential attacker? Hope these aren't stupid, Al.
Received on Thursday, 17 May 2001 17:09:14 UTC