[Ietf-dkim] Re: New I-D: A Deployment Profile for DKIM2 via Milter Interface

[Richard Clayton <[e-mail]>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

In message <[e-mail]>, Vittorio
<[e-mail]> writes

>Attribution of body modifications is already provided by the signature 
>chain: a change in bh= between two consecutive signatures identifies the 
>signing domain at the modifying hop. This is attribution: who modified 
>the body, with cryptographic proof. 

You don't mean "attribution" and you do not mean "proof"

     Honest Abe creates an email with body hash abc1

     Baddie Bob changes the body and claims the body hash is still abc1

     Collaborator Charlie copies the body and lies that the hash is def2

     Demonic Dave copies the body and truthfully says the hash is ghi3

     Eager Eddie gets the mail and attributes changes to C and D
     who are the two players who do NOT change the body

>What body recipes add is 
>reconstruction - what was changed, in sufficient detail to recover 
>previous body states. These are different properties with fundamentally 
>different deployment costs.

yes they are different. The people with running code don't seem to think
the difference in cost is fundamental.

>Your point about null recipes strengthens this separation. You write 
>that null recipes reintroduce trust and that this is a strong reason not 
>to support them. 

I don't think anyone thinks they should be "supported" absent a
contractual relationship with the entity being trusted.

>I agree with the analysis. But the consequence is that 
>every node in the delivery chain must produce a valid body recipe:

only if they change the body

> every 
>security gateway, antivirus engine, DLP system, mailing list manager. If 
>any single node fails to produce a recipe, reconstruction breaks for 
>every prior hop. 

correct

>This is a deployment requirement qualitatively 
>different from anything DKIM1, ARC or envelope binding demands

ARC requires you to trust random strangers ... with no ability to check
their working. That's why it has not been widely adopted

- -- 
richard @ highwayman . com                       "Nothing seems the same
                          Still you never see the change from day to day
                                And no-one notices the customs slip away"

-----BEGIN PGP SIGNATURE-----
Version: PGPsdk version 1.7.1

iQA/AwUBaeGDt2HfC/FfW545EQI7HgCgq1Rs4vVR9Co9HpCgMgbnjxFn8p0AnjVc
NY4f6nW89t0Dx284J7fyy+mc
=4Gjd
-----END PGP SIGNATURE-----