Re: Towards secure XML (sXML)

At 15:38 10/24/2000 +0200, P Brandt wrote:
>For some years we have been addressing the subject of XML and cryptography 
>at the Netherlands Organisation for Applied Scientific Research (TNO). 
>Please find attached a document describing in a nutshell some of our ideas 
>on how we would like to see secure XML to develop. We think the current 
>proposals could aid from this. If sufficiently relevant to the XML 
>encryption group we'd like to support its further elaboration.

Paul and Frederik,

Thank you for forwarding your note [1]. The introduction of the term 
"secure-XML" (of which XML Encryption is but a single piece among others) is 
one way to approach the larger problem; a few comments/questions follow:

>1.4 Authentication
>...
>The authentication aspect has been subject to study already by the Digital 
>Signature Working Group. Strictly speaking, digital signatures guarantees 
>two security assets: the authentication as well as the integrity. Already 
>some discussions have taken place within the XML encryption group regarding 
>this issue. In our believe, sXML should address the authentication issue as 
>a requirement however this can be done by incorporating the DigSign 
>recommendation.

As an aside, the sort of authentication XML Signature provides is message
authentication, and/or signer authentication where signer authentication 
means "key authentication" and not necessarily identity. XML Signature 
itself doesn't get into identity, but can work with applications (PKIX, PGP) 
that do.

...

Otherwise, reading your note and much of the recent discussion has prompted 
me to get close (but not quite) to the options I think we have with respect 
to the design:

UPON ENCRYPTION

1 require that the encryption not invalidate the target schema.
2 instances once modified need be only well formed.
3 require that the modified instance also have a modified schema.

UPON DECYRPTION

1 require that the decryption not invalidate the target schema.
2 well formed instances, when decrypted, yield the original valid document 
iff it was valid
3 require that the decrypted version also will be provided with its original 
schema.

I know that's not useful ... still fuzzy ... I'll keep thinking ... <smile>

[1] 
http://lists.w3.org/Archives/Public/xml-encryption/2000Oct/att-0016/02-Discussion_paper_sXML.doc

__
Joseph Reagle Jr.
W3C Policy Analyst                mailto:reagle@w3.org
IETF/W3C XML-Signature Co-Chair   http://www.w3.org/People/Reagle/

Received on Friday, 27 October 2000 19:17:43 UTC