AW: Header Protection for S/MIME

[Jörg Schwenk <[e-mail]>]

Hi,

to start the discussion on our draft [1], I'd like to explain the idea and
the advantages it has (in our opinion) over the solution proposed in RFC
3851.

Idea

The idea is to put Information about the header in a CMS hashed subpacket.
Legacy mail clients may simply ignore this subpacket (i.e. they don't
recompute the hash value contained in this packet). Conforming mail clients
compute an additional hash value over some normalized header fields, and
include this hash value in CMS signature verification.

Advantages

1. Backward compatibility: Legacy mail clients are not affected by the
introduction of an additional signed subpacket. (This is in contrast to RFC
3851, where legacy clients will display only the outer header lines, without
check.)

2. Flexibility: Any combination of header lines can be protected. This may
range from introducing a single hashed subpacket for the most important
header lines (e.g. From, Sender, To, CC, Date, Subject) to separate hashed
subpackets for each line. Conforming clients may thus detect changes in sets
of lines, or single lines.

3. Easy implementation: If a header line has been changed, a warning can be
displayed (e.g. display the line in red).

4. Support for mailing lists: If a mail list agent changes the To header,
and if To was protected by its own hashed subpacket, the mail client will
display that this field has been changed.

I think our draft still needs a lot of refinement, and we will be grateful
for comments from this list.

We have implemented a Java client as a proof-of-concept, and a thunderbird
implementation is underway.

[1] http://www.ietf.org/internet-drafts/draft-liao-smimeheaderprotect-00.txt

Greeting

Joerg
www.nds.rub.de

________________________________________
Von: Russ Housley [mailto:[e-mail]] 
Gesendet: Donnerstag, 11. Oktober 2007 18:13
An: [e-mail]
Cc: [e-mail]; [e-mail]
Betreff: Re: Header Protection for S/MIME

I have not seen any discussion of this document on this list.  It is
proposing a very different approach to a problem that was discussed on this
mail list.  The current MSG specification includes a very different solution
to this problem.

We should be talking about this proposal ....

Russ

At 09:29 AM 10/1/2007, Turner, Sean P. wrote:


The authors of the following draft wanted me to bring their draft to your
attention: 

http://www.ietf.org/internet-drafts/draft-liao-smimeheaderprotect-00.txt 

spt