pem-dev
[Top] [All Lists]

Re: Nested PEM Messages

1993-12-29 23:50:00
-----BEGIN PRIVACY-ENHANCED MESSAGE-----
Proc-Type: 4,MIC-CLEAR
Content-Domain: RFC822
Originator-ID-Asymmetric: MFMxCzAJBgNVBAYTAlVTMQswCQYDVQQIEwJNRDE
 kMCIGA1UEChMbVHJ1c3RlZCBJbmZvcm1hdGlvbiBTeXN0ZW1zMREwDwYDVQQLEwh
 HbGVud29vZA==,03
MIC-Info: RSA-MD5,RSA,ZvYI9yeICrn3g+LSK4KCUG0CS5hP1FDubFaiq86wMiU
 40ai2exB4kPKvh1MbB0BDetCwWv/0XkbSAWEn23xEm8BKb2i9g60E8omm8Fy0D/N
 mZIaeIXY3ktY11JnF0A4R

When TIS/PEM is invoked under MH, it "de-enhances" a PEM message that
is (usually) in the user's inbox and puts the result back into the
user's inbox.  If that message has a PEM body within it, the user can
invoke PEM processing again.

In the integration of MIME and PEM we're working on, one of the modes
will be to save the name of the sender and other information, e.g. how
and when the signature was checked, as a separate but related body
part.  This could be done recursively, but it's messy if it's desired
that annotations be recomputable and that signatures apply to
uninterpreted body parts.

        Let S(u,x) stand for a message with text x signed by user u.
        Let A(u,x,t) stand for a de-enhanced message with information
        t generated by the "de-enhancement" process, e.g. the
        distinguished name of the sender, time stamp information, etc.
        An implicit constraint is that t is completely redundant.  At
        any time, t can be discarded and regenerated.

        Suppose the user receives S(u1,S(u2,m)).  If only one step is
        applied, the result is A(u1,S(u2,m),t1).  This is fine.
        However, if we also dig into nested messages, we probably want
        something like A(u1,A(u2,m,t2),t1).  However, the annotation
        t1 properly applies to S(u2,m), not to A(u2,m,t2), so
        A(u1,A(u2,m,t2),t1) is not correct.

One method of recursively processing a message which keeps annotations
and leaves the original message intact requires mirroring the message
structure in an auxiliary structure.  This requires at least some
replication of information, and may result in exponential increase in
message length.  This approach would likely make the current PEM, MIME
and X.400 formats look efficient.

Steve
-----END PRIVACY-ENHANCED MESSAGE-----

<Prev in Thread] Current Thread [Next in Thread>