[LINK] Demolition Job on Digital Signatures
Thu, 9 Nov 2000 13:20:45 +1100 (EST)
On Thu, 9 Nov 2000, Roger Clarke wrote:
> I've finally got round to codifying the problems with conventional public
> key infrastructures, i.e. those based on X.509 certificates.
> The submission deadline for the conference I'm targeting is 24 hours away,
> so brisk, constructively negative responses would be greatly appreciated.
Is a non-constructive positive response is out of the question?
Your paper echos many of my own concerns regarding digital certs
and identity authentication. The fact that you recieve a document
digitally signed with Rick W's key in NO WAY GUARANTEES that
Rick W signed the document. It only indicate that my key was used.
The decoupling of personal identity with key identity, added to the
inherent insecurity of machines holding these keys, and those machines
being (semi-)permanently connected to the Internet makes digital
signing using X.509 unreliable and at worst misleading in its false
sense of secure identification.
You've hit several nails on the head here.
My own postscript: I am working with various e-commerce (B2B) protocols,
some of which encourage or even require the use of digital certs. Without
fail, each customer I have hooked up to our supplier system (regardless
of protocol) has abrogated using digital certs in preference for something
simpler, like username/password over SSL. And not for the reasons
Roger cites in his paper ... its just too damn complicated for most
non-techies to obtain and administer digital certs with present technology.
And I detect a definite fear factor at work here as well ... a healthy
fear of the unknown.
Rick Welykochy || Praxis Services
"Tired of being a crash test dummy for Microsoft? Try Linux"