tem, which is already in place, can be re-used for personal content. Second, the use of
a DRM system for protecting personal content is thought to create for users a more
positive view of DRM in general, and this can contribute to the acceptance by users of
controlled distribution of commercial content.
The original DRM system involves an authorization hierarchy implemented by
means of competent authorities that create and sign digital certificates. This authoriza-
tion hierarchy is modified in the extended DRM system to accommodate the fact that
consumers can now be content providers as well. For this reason, a new digital certifi-
cate (i.e., a “certificate of ownership”) is introduced which establishes a secure link
between a user and his personal content.
The fact that consumers become content providers has security implications as
well. Consumers are now able to control the usage of their personal content in the
system, but this also opens the door for potential misuse of commercial content (e.g.,
its illegal introduction in the system as personal content). To prevent such a threat, the
extended DRM system requires that users register their content with a competent au-
thority, at which point they obtain the certificate of ownership and may have their
content marked with their identity, but only after the identity of the content itself has
been checked and the content certified as new.
The extended DRM system can provide further user privacy by providing users
with the possibility of private ownership of content with private and controlled multi-
ple ownership. This means that users are able to register their personal content under
pseudonyms, with unlinkability of pseudonyms also supported (i.e., a unique pseudo-
nym per content item). Moreover, multiple users may own a content item, with their
privacy protected in two ways: (i) pseudonyms, a different one for each user, can be
used for content registration, and (ii) transfer of ownership of their content must be
decided jointly by all owners.
Privacy of content ownership can be achieved as described above, except towards
the registration authority which always keeps a record of the original user identifier
and all the content registered under the corresponding pseudonyms. This is done in
order to enforce accountability in the system but may be seen as a downside of the
system. This lack of privacy can be alleviated by means of a mechanism of distribu-
tion of trust. In this case, the original user identifier can be replaced by temporary
identifiers which are then used by the registration authority. The temporary identifiers
are, in their turn, generated by another trusted third party which must then (to enforce
accountability) keep a record of the user’s real identity. As long as the authorities do
not collude, the association between users and their personal content is not known by
any of the parties in the system. Of course, trust can be further distributed to diminish
the possibility of collusions between authorities.
While the mechanism described above would increase users’ privacy, it would also
make the technological solution more complex, certainly from the architectural point
of view. This trade-off between system’s privacy provision and system’s complexity is
encountered often, mainly in systems with the strong requirement that security levels
be preserved after addition of privacy enhancements. This is certainly the case for the
original system considered, i.e., a DRM system for the protection of commercial con-
tent. Its extension to protect personal content, as well as further extensions to provide
user’s privacy, must include mechanisms to ensure user accountability, otherwise
commercial content providers will certainly object to such extensions. Therefore,
184