Controlled Sharing of Personal Content using Digital
Rights Management
Claudine Conrado
1
, Milan Petkovic
1
, Michiel van der Veen
1
, Wytse van der Velde
2
Department of Information and System Security, Philips Research, P.O. box WY 71,
5656 AA Eindhoven, The Netherlands
Abstract. This paper describes a system which allows the controlled distribu-
tion of personal digital content by users. The system extends an existing Digital
Rights Management system that protects commercial copyrighted content by es-
sentially allowing users to become content providers. This fact however makes
the system vulnerable to the illegal content distribution, i.e., distribution by us-
ers who do not own the content. To solve this problem a solution is proposed
which involves the compulsory registration with a trusted authority of a user’s
personal content. During registration, content identity is initially checked to
verify whether the content is new. If it is, the association between user identity
and content is securely recorded by the authority, with users also having the
possibility to remain anonymous towards any other party. In this way, the
trusted authority can always verify personal content ownership. Moreover, in
case the initial content identification fails and content is illegally registered, the
authority can ensure user accountability.
1 Introduction
Recent developments in digital technologies, along with increasingly interconnected
high-speed networks and the decrease in prices for high-performance digital devices,
have established digital content distribution as one of the fastest emerging activities
nowadays. This trend of digital content distribution gives great opportunities for
commercial content providers but also poses threats as digital content can be very
easily illegally copied and distributed. Therefore, commercial content providers need
technologies, accompanied by legislation, that can protect digital content from illegal
use. Digital Rights Management (DRM) is a collection of technologies that provides
content protection by enforcing the use of digital content according to granted rights.
It enables content owners and content providers to protect their copyrights and main-
tain control over distribution of and access to content.
In parallel to the introduction of multimedia download services, there is also a clear
increase in production of personal digital information by consumers including digital
photos, home-video, text, and so on. This is inspired by the ever decreasing prices of
digital cameras, camcorders, mobile phones, storage devices, as well as cheaper high-
Conrado C., Petkovic M., van der Veen M. and van der Velde W. (2005).
Controlled Sharing of Personal Content using Digital Rights Management.
In Proceedings of the 3rd International Workshop on Security in Information Systems, pages 173-185
DOI: 10.5220/0002573101730185
Copyright
c
SciTePress
speed Internet. As a consequence, consumers have to deal with an ever growing
amount of personal digital data, alongside the downloaded commercial content. Some
of this personal content might be highly confidential and in need of protection, there-
fore users may want to share it in a controlled way. This means that the content owner
should be able to control the use of his content by other users with whom he shares it.
For instance, the owner could specify rights such as play, play once and copy, which
are enforced when those users access the content. More importantly, the owner could
control further distribution of his content by preventing the users with whom he shares
the content from spreading it further. However, in contrast to a lot of effort which has
been put into the protection of copyrighted commercial content in different DRM
systems currently available, controlled sharing of personal content is often ignored.
In this paper, an approach to controlled sharing of personal content is described
which reuses the concepts defined in DRM systems used by the commercial content
owners. However, when personal content and commercial content are managed to-
gether in one system, the clear line between producers and consumers of content (or
content providers and users) fades. Every user of the system is a potential content
provider and is able to create and introduce new content into the system. As a conse-
quence, numerous new issues arise such as the requirement that the security of the
system should remain the same as before the introduction of personal content. This
means that the system should prevent illegal distribution of commercial content as
well as personal content. Further requirements are that content owners should be able
to protect their privacy and the security of their content, even when jointly owning
content. Therefore, while extending the DRM system to support controlled sharing of
personal content, focus is put on preserving the system security.
The remainder of the paper is organized as follows. In Section 2, the related work
is reviewed. Section 3 discusses a solution that extends an existing DRM system for
commercial content to support controlled sharing of personal content. In Section 4, a
description of a method that prevents illegal sharing of commercial content and its
security’s aspects is given. Section 5 describes an extension of the system that en-
hances user privacy and allows private multiple ownership of the content. Finally,
Section 6 draws conclusions.
2 Related Work
Sharing of digital personal content, especially over the Internet, is nowadays wide-
spread. For instance, P2P distribution networks have become one of the most used
technologies for sharing content between people, which is attested by the popularity of
applications such as KaZaa. However, these applications do not provide any control
over the distribution of content and therefore allow illegal distribution of commercial
content, in particular copyrighted music and movies.
From the commercial content industry’s point of view, illegal sharing of protected
commercial content is one of the major downsides of the P2P technology. However,
P2P networks do have the potential to provide an industry-compliant distribution
model. The success of iTunes [1] from Apple provides some evidence that there are
possibilities of using copyright-compliant systems in a profitable way, provided that
174
the incentives for using such a system are high enough. In the case of iTunes, these
incentives are quality, availability and price of music.
Other proposals for providing such incentives have been presented, e.g., the Mu-
sic2Share system [2]. It uses quality control of music and combines this with relatively
free usage restrictions and an easy payment structure to make it interesting for users.
Another proposal by Grimm and Nützel [3] combines the idea of DRM and P2P
file sharing with the idea that a user who has paid for content can redistribute this
content against payment. Users that do not pay are unable to redistribute content, thus
providing an incentive for payment.
In this paper, a different way to create an incentive is proposed: users of a commer-
cial content protection system can utilize the same protection mechanisms which are
used for commercial content also for their own personal content. In this way, the DRM
system, instead of only restricting users, also helps those users by protecting their
personal content. This presumably gives a much more positive image to DRM as a
technology which can be in fact beneficial to users. Moreover, it gives an incentive to
users to support the system and prevent that its security be compromised in any way.
3 A DRM System for Personal and Commercial Content
In this section, a DRM system for controlled distribution and usage of commercial as
well as personal content is presented. The basics of the system, its components and
settings are described in the next section, while section 3.2 explains how the system is
used for content protection and sharing.
3.1 System components and authorization hierarchy
The system is an extension of a DRM framework [4] for protection of commercial
content. In the extended DRM system, users are able to protect and controllably share
their personal content. This effectively means that the user who is the owner of the
content takes over a role of content and license provider and therefore becomes in-
volved in the content and license creation as well as content protection tasks. Consid-
ering personal content, owners will most likely be sharing content with other users, so
person-based DRM model [5], in which rights to access content are granted to persons
rather than devices, is used as a basis for the presently described DRM system.
While describing the system, components and properties are highlighted which are
relevant for the protection and sharing of personal content. Most of these system com-
ponents are also part of the original DRM system for protection of commercial content
and are described below:
(i) personalized smartcard: a device used for user identification. The smartcard
contains a private key which corresponds to the public key of the user. The
public key is certified by some Trusted Third Party (TTP). Therefore, the TTP
knows the link between the user’s identity and his public key.
(ii) compliant device: a device that behaves according to the DRM rules. It can
identify a user by means of a personalized smartcard. The compliant devices
175
provide a secure and robust computing platform for the DRM operations and
offer a secure storage for keys and state information used in the system.
(iii) content container: personal digital content is stored in encrypted form together
with a unique content identifier (the “content container” abstraction) and can
be retrieved from anywhere in the network.
(iv) Content Right: a certificate issued by the content provider containing the con-
tent identifier, a content encryption key (CEK), the User Right Authority’s
public key and the content provider’s signature. The Content Right is securely
stored in the compliant device or securely obtained from a content provider.
(v) User Right: a certificate issued by the User Right Authority authorizing a per-
son to use content according to granted rights. It consists of content identifier,
user identifier, rights expression and the User Right Authority’s signature. User
Rights may reside anywhere in the system and are not stored in a secure way
1
.
(vi) Content ID Certificate: a certificate that creates a secure link between content,
its identifier and the content owner, who becomes also content provider. This
certificate is needed for personal content, since the content provider may be
any user in the system and needs, thus, to be explicitly linked to the content.
The certificates used in the system are issued by different authorities, which organ-
ize an authorization hierarchy depicted in Figure 1.
Fig. 1. Certificate Authorization Hierarchy
The System Authority is the root of the hierarchy. Each compliant device has a
built-in key of this authority. The Service Provider Authority (SPA) authorizes differ-
ent content providers (users) by certifying their public keys and linking them with
content items by means of the Content ID Certificate. Content providers are allowed
to issue Content Rights. A Content Right contains in turn the public key of the User
1
Content Right and User Right may be merged in one license as in the OMA DRM [9].
176
Right Authority. This construction allows content providers to delegate the issuing of
User Rights. The User Right Authority issues User Rights for a certain piece of con-
tent. The Device Authority authorizes Device ID Authorities, e.g. device manufac-
tures, to issue Device ID Certificates that give a unique identity to a device. The User
ID Authority issues User ID Certificates. This certificate typically contains user iden-
tity information and a public key that identifies a user within the system.
3.2 Content protection and sharing
During the content protection and sharing procedures, a number of different transac-
tions, schematically depicted in Figure 2, are performed within the system. These
transactions are described below
2
, where references to the numbered links in Figure 2
are made at the appropriate points.
When a user Alice wants to protect a personal content item, in order to be able later
on to share it in a controlled way, she registers the content with the SPA (content
registration is discussed in the next section) and obtains a Content ID Certificate. The
certificate consists of the fields Content ID, User ID, Content Fingerprint and Signa-
ture SPA. The Content ID field refers to the content itself. The User ID field contains
the public key of Alice, who is then considered the owner of the content (as well as
content provider). The Content Fingerprint is a short “summary” of the multimedia
content which can uniquely identify that content. It usually consists of robust features
extracted from the content [6]. Finally, the Signature SPA is the digital signature of
the SPA on the certificate. The next step in the content protection process is creation
of the Content Right together with encryption of the content. Before allowing Alice to
create a Content Right, a compliant device authenticates her (step 1 in Fig. 1) and
checks whether she is authorized to do so by inspecting the Content ID Certificate,
which must contain Alice’s ID (her public key). If so, the compliant device proceeds
with the creation of the Content Right certificate (step 2 in Fig. 1) by randomly choos-
ing a symmetric key (CEK), which is used to encrypt the content, and placing the key
in the certificate. Additional information in the Content Right certificate includes the
Content ID and the public key of the User Right Authority. Typically, this authority
will be the content owner Alice, but she may also create a Content Right for other
users allowing them to sign User Rights (i.e. to share the content). Finally, Alice signs
the Content Right.
2
The described transactions for content protection refer specifically to personal content: a user
can act as a content provider, therefore this must be certified by means of the Content ID
Certificate. The described transactions for content sharing, on the other hand, are essentially
the same for both, personal and commercial content, except that in the latter a (recognized)
commercial content provider takes the role of the user.
177
Fig. 2. Representation of the various interactions in the DRM system
If Alice wants to share such protected content item with another user Bob, she cre-
ates a User Right certificate that includes the public key of Bob and sends the secure
content container together with the necessary certificates (User Right and Content ID
Certificate) to Bob (step 3 in Fig. 1). The Content Right certificate is also needed but
it should be securely sent directly to the compliant device that will be handling the
content access action with Bob, since that certificate has the CEK, which can be used
to decrypt the content. When Bob wants to access the content, he must have a valid
User ID Certificate. He authenticates to a compliant device using his smartcard (step 4
Fig. 1) and presents the obtained certificates from Alice together with the secure con-
tent container. The compliant device must check the validity of the certificates before
allowing access to the content (step 5 Fig 1.). Moreover, the device will compare the
User ID from the User Right with the User ID in the User ID Certificate. It also has to
verify that all certificates are used for the content whose identifier is in the Content ID
Certificate. To do so, the compliant device obtains the CEK from the Content Right
certificate, decrypts the content, calculates the fingerprint of that content and finally
matches the fingerprint of the content with the fingerprint in the Content ID Certifi-
cate. If the fingerprints do not match, the device will detect a content substitution
attack, which means that the original encrypted content in the content container has
been replaced with a different content, but assigned the same identifier. The device in
this case will not allow Bob to access the content.
4 Registration of Personal Content
Any content which is to be protected under a DRM system must be recognizable
within that system. In the case of commercial content, typically each content item is
assigned a unique content identifier within that specific DRM system. This assignment
can be thought of as the registration of the content item in the DRM system.
The DRM system described in the previous section is used to protect both commer-
cial and personal content. The assignment of a unique content identifier to every piece
178
of content that enters the system is still necessary. However, as now every user can
become a content provider and sign certificates in such a system, there is a danger
regarding illegal distribution of commercial content. The problem exists because there
are no explicit measures to prevent commercial content from being distributed as
personal content by users of the DRM system. Without preventive measures, a mali-
cious user who has obtained an illegal copy of commercial content is able to distribute
that content within the DRM system as his own personal content.
A solution to the problem above is to introduce a secure registration procedure that
not only assigns a unique identifier to content, but also verifies the identity of content
by using fingerprint technology [6]. The registration procedure as well as its security
aspects are discussed below.
4.1 The Registration Steps
To introduce a new content item into the system, the user Alice has to follow a number
of steps which are shown in Figure 3 and are described below.
Fig. 3. The Registration Phase
Step 1. Alice imports to her compliant device a content item that she wants to intro-
duce. The device calculates the fingerprint of the content item.
Step 2. The device provides the SPA with the fingerprint of the content item and the
public key of Alice.
Step 3. The SPA matches the fingerprint of Alice’s content against a database of
fingerprints from known commercial content and already registered personal
content.
179
Step 4. If there is a match the SPA refuses to register the content item. Otherwise, the
SPA proceeds by generating a Content ID Certificate and a database entry,
with a time stamp on it, linking Alice’s ID to the content (i.e., to its finger-
print). This allows the SPA to prevent other users from trying to later on reg-
ister Alice’s content as their own. Optionally, a watermark identifier and a
watermarking key [6] are also generated by the SPA to allow the possibility
of “forensic tracking” outside the DRM system, as discussed below.
Step 5. The SPA sends the Content ID Certificate (and the watermark identifier and
watermarking key in case watermarks are used) to the compliant device.
Step 6. Finally, all the necessary certificates are created so as to make the content
suitable for distribution using the DRM system. When watermarks are used,
the compliant device also watermarks the content with the watermark ID us-
ing the watermarking key.
4.2 Security Analysis
The threat model used for the present security analysis is as follows. A casual-copying
model is assumed as described in [7]. This means that the security of the system is not
concerned with widespread copying and distribution by organized groups of criminals,
but with small-scale and disorganized copying amongst small groups of ordinary us-
ers. The security of the system tries to frustrate and create barriers for would-be in-
fringers. In the threat model, it is assumed that if the barriers for illegal distribution
are high enough and there is an incentive for users to behave in accordance with the
system, then most of the users will behave in such a way. Therefore, if only a few
users manage to distribute illegal content within the system, the system itself does not
fail regarding the threat model. With such a threat model in mind, specific system
security aspects are discussed below.
Fingerprint Check and DRM Infrastructure. The goal of fingerprint identification
at the beginning of the registration procedure is to prevent malicious users from
registering commercial and pre-registered personal content as their personal content. It
is assumed that the SPA has a database with the fingerprints of all existing commercial
content as well as pre-registered personal content. In a more realistic setting, the SPA
may use a distributed system in which the query to match a given fingerprint is
forwarded to other parties with their databases. The management of such databases is,
however, outside the scope of this paper.
After fingerprint check has been performed, content is rejected in case a match is
found, otherwise it is accepted for registration in the system. However, the fingerprint
identification may fail, for instance, if the checked fingerprint databases are incom-
plete and therefore unable to recognize all registered content, or if the malicious user
manages to tamper with the content and changes its fingerprint so that it cannot be
recognized. In the latter case, however, the robustness of the fingerprint should be
sufficient to substantially degrade the quality of the changed content. In any of these
cases, the consequence of fingerprint identification failure is that content is success-
180
fully and illegally registered and can be distributed within the DRM system as belong-
ing to the malicious user.
The very infrastructure of the DRM system, however, can be used as a mechanism
against illegal registration of content. This infrastructure is based on digital certificates
issued by competent authorities, as depicted in Figure 1. In order for a registered con-
tent item to be distributed within the DRM system, certificates must be issued and
presented to compliant devices for verification. In particular, a Content ID Certificate
is issued by the SPA which contains the User ID (public key) of the content introdu-
cer. Moreover, this public key is also used to sign Content Rights and possibly User
Rights. Therefore, if illegal content is detected within the DRM system, the original
content introducer can be found via his public key and the corresponding User ID
Certificate.
Watermark Extraction. A further measure which can be used to increase the
system’s security is the use of watermarks. During the registration phase, the content
may be optionally watermarked by the compliant device with a watermark ID which is
created and sent to the device by the SPA. This acts as a pointer to a user’s ID in the
SPA’s database. The primary aim of such a watermark is to serve as a “stamp” of
ownership which the content item carries wherever it goes. Therefore, in case of a
dispute, for instance, over content ownership, it can always be determined by means of
the watermark.
The watermark also allows the possibility of the so-called “forensic tracking” out-
side the DRM system. This covers the situation in which illegal content has been in-
troduced into the DRM system by a malicious user and has been, later on, brought
outside the DRM system. In this case, the infrastructure of the DRM system can no
longer be used to identify the malicious user. Moreover, tampering with the content
may lead to changes in its fingerprint in a way that it can no longer be recognized. By
means of the watermark ID, however, the malicious user’s identity can be recovered
from the SPA’s database.
Illegal content in this context may be (i) commercial content, which typically is not
originally watermarked but has its ownership certified by other means, and (ii) other
user’s registered personal content, which will in this case include an additional wa-
termark ID pointing to the original owner’s ID (and therefore with an earlier time
stamp on it). In either of these cases, ownership had been established prior to the reg-
istration performed by the malicious user, so malicious behaviour, i.e. the illegal regis-
tration of content, can be proved.
Forensic tracking outside the DRM system is performed with participation of the
SPA (for the watermark extraction and database check) as well as participation of the
User ID Authority (which knows the real identity of the user whose public key is in
the SPA’s database).
5 Private and Controlled Ownership Sharing
In order to register a content item, the user has to authenticate with the SPA, so he is
not (and should not be) anonymous towards the SPA. However, he does not need to
181
identify himself to any other party in the system. In particular, a user may require that
he cannot be linked to a given content (anonymous publishing) nor linked to other
users for whom he creates User Rights. Moreover, a user may require that all his con-
tent items cannot be linked to one another and to his identity. These are user privacy
requirements which can be fulfilled by means of content registration with user pseu-
donyms. Additionally, if a different pseudonym is used for each different piece of
registered content, unlinkability of content items can also be achieved.
A further requirement is that shared ownership of personal content amongst multi-
ple users be possible. For instance, a family vacation movie in principle belongs to the
group of family members. In this case, the whole group participates in the registration
of the content and therefore acquires all rights implied by content ownership. It is
important to notice that ownership sharing, which is established at the registration
phase and addressed in this section, is distinct from the sharing of the content itself,
which is performed via the DRM system as described in Section 3.2. The rights im-
plied by content ownership include, e.g., access to the decrypted content as well as
power to determine who can distribute the content as User Right Authority. But as
personal content is likely to contain private information about its owners, an addi-
tional privacy requirement is that further sharing of ownership with outsiders (who
would then acquire owners’ rights) be jointly decided by all owners. This requirement
aims to limit the dissemination of personal content in a large and uncontrollable scale
since, via a content ownership sharing chain, the content may get out of the control of
the DRM system.
5.1 Registering Content with Controlled Anonymity
Pseudonyms generated from a user’s public key can be easily obtained from the SPA
in the following way. Assuming that the public/private key pair is built according to
the Diffie-Hellman key agreement, let x be the user’s private key and h=g
x
be the
corresponding public key, where g is a group generator (see [8] for further details). To
generate a pseudonym h’, the SPA generates a random value a and computes the new
public key h’=h
a
whose corresponding private key is x’=xa. The value a is then se-
curely sent to the user, so that he is able to calculate his new key pair (x’, h’), with no
other party learning his pseudonym.
Content can now be registered under the user’s pseudonym h’, which no party (ex-
cept the SPA) can link to the original public key h. For any new piece of content, a
new pseudonym can be generated in the same way and used for registration. When
ownership of a given content is shared between N owners, the procedure above is
repeated for all owners. Upon registration, all owners i (where i
{1, 2,…, N}) obtain
from the SPA distinct pseudonyms which are unlinkable to one another, to the owners’
real identities and to further pseudonyms the owners may get. The registration proce-
dure proceeds as described in the previous section with a few modifications. After
content fingerprint checking, the SPA generates random values a
i
and computes the
pseudonyms h
i
’=h
i
ai
. Each value a
i
is sent to (and only to) owner i. The Content ID
Certificate now comprises a unique content identifier, the content fingerprint and the
new pseudonyms h
i
of each registered owner. This certificate serves also to ensure
182
that all the pseudonyms h
i
are constructed in the correct way. Moreover, information
stored in the SPA’s database now must include the original identifiers h
i
as well as the
generated random values a
i
for each of the owners.
Because the SPA stores the values h
i
and a
i
for all owners of a given content, any
pseudonym and/or any content can be traced back to any of the owners’ real identities.
This provides revocable anonymity which is important to provide accountability in the
system. For instance, illegal import of content into the DRM system can be traced
back to the culprits, even if they are normally anonymous in the system.
5.2 Controlled Ownership Sharing
A further privacy issue is how to control the further sharing of ownership. This sharing
should be allowed only if jointly decided by all owners (in which case the registration
process is carried out again with an extra owner added to the data). Sharing, however,
may happen without the approval of all owners. This can happen via the disclosure by
one of the owners, say i (the “leaker”), to an outsider of the pseudonymous private key
x
i
. With this key, the outsider has all rights of the leaker in the DRM system. And
since the outsider does not learn anything about the leaker’s original private key x
i
(given that x
i
is randomly related to x
i
), a deterrent mechanism is needed to stop such
potential leakers. Such a mechanism is described below.
The SPA is set up such that it will reveal the value a
i
(of any of the owners i) to any
party who comes to it and proves knowledge of the private key x
i
corresponding to
the pseudonym h
i
. The party may do so completely anonymously. The idea behind
this is that only one of the two, the owner i himself or an outsider who received the
private key x
i
from owner i, is able to authenticate with pseudonym h
i
. In the first
case, there are no consequences since owner i should in fact know a
i
3
. The second
case, however, has major implications since the outsider is able to learn the leaker’s
original private key by calculating x
i
= x
i
’/a
i
. Given that the outsider may considerably
profit from the knowledge of x
i
, with no liabilities on his part, he is likely to do so. A
further implication of the outsider’s action is the fact that the SPA is immediately
notified that owner i has leaked his private key to an outsider. The SPA can then warn
the other owners about the fact and, furthermore, can tell them who the leaker was.
6 Conclusion
This paper presents and discusses a system for the protection of users’ privacy when
sharing personal digital content with other users. The system allows the controlled
sharing of personal content, which is achieved by means of an extension of a DRM
system originally devised to protect commercial content. The idea behind extending
an existing DRM system to protect both types of content is two-fold. First, the whole
infrastructure (e.g., compliant devices and certificate authorities) of the existing sys-
3
Owner i may need to ask this value in case, e.g., he looses it. In this case, he must authenticate
with his original public key h
i
, in order to prove that the request is legitimate.
183
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
privacy protection has limitations in such systems. This means that it is unrealistic for
any user, specially the user who values the freedom of anonymous publishing above
the right to claim content ownership, to expect completely anonymity in the system.
References
1. Apple iTunes, http://www.apple.com/itunes.
2. T. Kalker, D. H. J. Epema, P. H. Hartel, R. L. Lagendijk, and M. van Steen, “Music2Share
- Copyright-Compliant Music Sharing in P2P Systems,” Proceedings of the IEEE, Vol. 92,
No. 6, June 2004.
3. R. Grimm and J. Nützel, “A friendly peer-to-peer file sharing system with profit but with-
out copy protection”, in Innovative Internet Computing Systems, Lecture Notes in Com-
puter Science, Springer-Verlag, vol. 2346, 2002, pp. 133–142.
4. S.A.F.A. van den Heuvel, W. Jonker, F.L.A.J. Kamperman, P.J. Lenoir, “Secure Content
Management In Authorised Domains”, IBC 2002, Amsterdam 2002.
5. C. Conrado, F. Kamperman, G.J. Schrijen and W. Jonker, “Privacy in an Identity-based
DRM System”, Proceedings of the 14
th
International Workshop on Databases and Expert
Systems Applications, Prague, Czech Republic, 2003.
6. M. van der Veen, A. Lemma, and T. Kalker, “Watermarking and Fingerprinting for Elec-
tronic Music Delivery System”, SPIE 2004, San Jose, USA.
7. Freedom to Tinker, March 2003. http://www.freedom-to-tinker.com/archives/000317.html.
8. A.J. Menezes, P.C. van Oorschot, S.A. Vanstone, Handbook of Applied Cryptography,
CRC Press, 1997.
9. DRM Specification V2.0, Open Mobile Alliance, OMA-DRM-DRM-V2_0-20040716-C,
2004.
185