- From: Tom Gindin <tgindin@us.ibm.com>
- Date: Wed, 5 Feb 2003 14:22:46 -0500
- To: Rich Salz <rsalz@datapower.com>
- Cc: Joseph Swaminathan <jswamina@cisco.com>, w3c-ietf-xmldsig@w3.org
Rich: My own position is a little different. I think that the presence of both KeyValue and X509SubjectName suggests that X509SubjectName is unreliable, and is likely (not just possibly) the aftereffect of an attack. On a similar subject, is it really a good idea for X509Certificate to be present with either X509SubjectName or X509IssuerSerial? If the RP application uses the smaller field for display purposes to avoid the complexity of parsing the certificate, it opens a similar attack to the KeyValue/SubjectName one, and if it doesn't the smaller field is pointless. Tom Gindin Rich Salz <rsalz@datapower.com> on 02/05/2003 01:33:17 PM To: Joseph Swaminathan <jswamina@cisco.com> cc: Tom Gindin/Watson/IBM@IBMUS, w3c-ietf-xmldsig@w3.org Subject: Re: X509 data element > My question is, if there is a content in the XML document we > cannot trust, then shouldnt we, not use it for any purpose. What > situation a data which can't be trusted be useful. Signature validation might be performed by a third-party service that has no knowledge of the signer identities; separating authentication from authorization. Perhaps it might help if you think of validation as a tri-state: trusted, untrusted, and indeterminate. <example removed> Your example can be summarized like this: the organization is using unsigned data in its operations, and that can be hacked. I agree. But that's irrelevant here. /r$
Received on Wednesday, 5 February 2003 14:24:29 UTC