Re: Protecting private key on a soft cert
From: saict (thurberk_at_cscsw.com)
Date: 21 Jun 2004 14:23:03 -0700
"Michel Gallant" <email@example.com> wrote in message news:<OsAMYMIVEHA.1292@TK2MSFTNGP10.phx.gbl>...
> "saict" <firstname.lastname@example.org> wrote in message
> > "Michel Gallant" <email@example.com> wrote in message
> > > CSP protection for MS providers in W2k+ is based on DPAPI.
> > > A fair level of details on how the various keys are derived based on
> > > logged on user is described here:
> > >
> > >
> > >
> > > - Mitch Gallant
> > > MVP Security
> > Thank you again, Mitch, for your valuable feedback. Would I be
> > correct in saying that the private key is, in the end, protected by
> > the user's login credential?
> If you don't provide extra "entropy" (i.e. "Strong Protection" option), than that is
> correct. Any process running under your loging credentials would have access to
> using the private key. That is why Strong Protection (based on another password-derived
> encryption) is important when importing (say from PKCS #12) private key to CSP.
> This is described fairly well in the dpapi article above.
> - Mitch
Yes, I understand that. I was just trying to get a feel for how
securely this was happening by default. A discussion at a PKI
conference had most of us thinking this was done by some sort of
'security by obscurity' scheme, so we were happy to get your answer.