- From: Richard D. Brown <rdbrown@GlobeSet.com>
- Date: Tue, 27 Apr 1999 13:07:45 -0500
- To: "'Phillip M Hallam-Baker'" <pbaker@verisign.com>, "'Bede McCall'" <bede@mitre.org>, <w3c-xml-sig-ws@w3.org>
Phill, Agreed that CMS shall refer to the IETF specification. However, a large majority of existing implementations are PKCS#7 and not CMS. So, I do not understand the argument developed previously on the list - I thought that support for CMS was motivated by the possibility to leverage existing implementations! Recall that CMS and PKCS#7 SignedData type are very similar, but there are not compatible even if you disregard CMS added functionality. Sincerely, Richard D. Brown > -----Original Message----- > From: Phillip M Hallam-Baker [mailto:pbaker@verisign.com] > Sent: Monday, April 26, 1999 12:02 PM > To: rdbrown@GlobeSet.com; 'Bede McCall'; w3c-xml-sig-ws@w3.org > Subject: RE: XML interface with URIs > > > > 1 - What do people refer to by CMS? CMS as specified by PKIX or > > PKCS#7 from > > RSA. > > CMS is the IETF interpretation of PKCS#7. At this point CMS is the > standard to reference. > > > 2 - CMS implementations usually require the > certificate-chain to be either > > refer to or pass as an argument. What is the impact on XML-DSIG > > implementation? Other crypto-algorithms require only the > private-key. > > I think as far as 'blobism' goes it is the detached signature > blob which > is of interest - everything within the signature envelope. > > PKI implementations require a certificate chain to authenticate a > signed object, at least according to PKI as we know it. Whether the > certificates are sent with the message, retrieved from a server > or directory there is a need to authenticate public keys in some > manner. > > I don't know of any PKI, including PGP which does not have such > a constraint. Certainly certificate chain transport is something > the XML spec has to address. It is not something which I would > want to insist on CMS to achieve however. Signature blobs stripped > of the cert chain achieve the backwards compatibility we need. > > > Also, we can make sure that the specification provides for > CMS without > > making CMS mandatory. Actually, I would certainly vote > against such a > > proposition. > > Votes? What votes? > > > Phill >
Received on Tuesday, 27 April 1999 14:07:32 UTC