Re: PGP and Compliance with SEC and Liability Rules

[Adam Back <aba@dcs.ex.ac.uk>]

Mark Grant <mark@unicorn.com> writes:
> Tim May quoted from macweek:
> >"The Gartner Group's Wheatman pointed out that PGP Policy Management Agent
> >allows corporatins for the first time to centralize control over
> >encryption: "For encryption to be accepted, IT had to gain control. This
> >isn't Big Brother; this is necessary to comply with liability laws and SEC
> >regulations.""
> 
> However, this doesn't seem to work, unless I'm mistaken about CMR
> enforcement and the SEC regulations. CMR will only allow the snoops 
> to read incoming email, not outgoing, and hence if Joe Blow at 
> Foo-Bah.com wants to send me some handy insider trading tips CMR will 
> not stop them. So this seems to be another justification for CMR 
> which just doesn't make sense.

I think that there is also a facility in the pgp5.5 for business
client to add yet another recipient: the sending companies snoop key.

I also got the impression that the policy enforcer could be set up to
bounce mail internally which did not have this extra recipient.

(Note, not having pgp5.5 for business to play with I am going on what
others have said.)

So now you've got mail headed out encrypted to 3 long term keys... if
the sender uses encrypt to self that will be 4 long term keys!  Then
the spooks in the sending country will want to be another recipient,
and spooks in the receiving country will also, bringing us to a total
of 6 long term keys.  Wew talk about security risks!

I thought it would be a good feature to put into the SMTP agent to
strip encrypt to self and recipients intended for internal snooping --
that would bring the recipients back down to 2 (or 4 with spooks).

Better still would be one recipient (as happens with adhoc local
escrow), and make that key a short term key which is burned after
expiry also.

Adam
-- 
Now officially an EAR violation...
Have *you* exported RSA today? --> http://www.dcs.ex.ac.uk/~aba/rsa/

print pack"C*",split/\D+/,`echo "16iII*o\U@{$/=$z;[(pop,pop,unpack"H*",<>
)]}\EsMsKsN0[lN*1lK[d2%Sa2/d0<X+d*lMLa^*lN%0]dsXx++lMlN/dsM0<J]dsJxp"|dc`