Abstract
Privacy of information is becoming more and more important as we start
trusting unknown computers, servers and organisations with more and more
of our personal information. We distribute our private information on an
ever-increasing number of computers daily, and we effectively give target organisations
carte blanche to do what they want with our private information
once they have collected it. We have only their privacy policy as a possible
safeguard against misuse of our private information. Thus far, no reliable
and practical method to enforce privacy has been discovered.
In this thesis we look at ways to enforce the privacy of information. In
order to do this, we first present a classification of private information based
on the purpose it is acquired for. This will then enable us to tailor protection
methods in such a way that the purpose the information is acquired for
can still be fulfilled. We propose three distinct methods to protect such
information.
The first method, that of nondisclosure, is where private information
is required not for the contents, but as input to verify calculations. We
shall present an encryption method to protect private information where
the private information consists of a set of numeric values S on which some
function G has to be applied and the result = G(S) has to be supplied to
a target organisation. The calculation of the result must be verifiable by
the target organisation, without disclosing S.
The second method, that of retaining control is a method by which we
can grant limited access to our private information, and thus enforce the
terms of privacy policies.
The final method we present is a conceptual method to extend P3P in
order to add more flexibility to the decision on whether or not a given item
of private information will be supplied to a target organisation by using the
Chinese Wall security policy. This will enable a user to not only define rules
as to which items of private information he would disclose, but also to define
what collection of private information any given organisation would be able
to build about him.
Olivier, M.S., Prof.