other music players is prohibited. The DRM system
proposed in (Messerges & Dabbish, 2003) allows
copying of licenses within a defined family domain.
The backup/restore functionality of all these
systems have several shortcomings. The threshold
limit might through unfortunate circumstances be
reached/exceeded, or it might for good reasons be
set to zero (backup/restore is disabled) by a content
owner who does not wish to open up for any
possibility of potential illicit copying, e.g. because
the content has a high commercial value. Forming
and maintaining device domains will have an
administrative cost and might be a hassle-factor for
the user – and what will the rules be for allowing
new devices into a domain? If a domain cannot be
established and the device gets lost or becomes
obsolete, restore can not take place. If the license
contains stateful information, restore cannot take
place either. In sum, there are still quite a few
pitfalls which may hinder effective backup.
4.3 License Transfer
This section proposes a license transfer mechanism
and protocol which meets some of the challenges
pointed out above. It reconciles the apparent
conflicting requirements for maintaining license
backups while preventing license duplication.
Consider a single-user, unlimited license which
we may label L. For simplicity we ignore the past, so
the license will be valid from now and onwards, i.e.
within the timeframe [t
1
, unlimited〉, where t
1
denotes the current time.
The basic underlying idea is to consecutively
subtract sub-licenses from L: Initially a new license
L
1
is subtracted from L. L
1
is valid during an
appropriately short time interval
Δ
t=t
2
-t
1
, i.e. during
the timeframe [t
1
, t
2
〉. At the same time, the validity
of L is reduced accordingly, and is now valid for the
remaining timeframe [t
2
, unlimited〉. License L is
kept on an “online backup server”. L
1
on the other
hand is transferred securely via a piracy-proof
channel (i.e. a channel which does not allow
duplication or “double-spending” of the license L
1
)
to its destination device, for example the SIM card
of a portable phone, where it is consumed. Once L
1
expires, or preferably a short time before, either the
destination device requests a new license from the
backup server (license pull), or the backup server
automatically transmits a new license L
2
(license
push). This license L
2
is, and subsequent licenses L
3
,
L
4
, … are extracted from L in the same manner as
L
1
.
The same mechanism can be applied to other
license types, e.g. to time-restricted licenses, where
the license L now has a timeframe [t
1
, t
n
] for two
specific points in time (e.g. from 1.1. to 31.12.). A
time-restricted license could also be a duration
Δ
t
(e.g.
Δ
t = 30d) which is converted into a fixed
timeframe [t
1
, t
n
] upon first usage. For counter-based
licenses, the counter (e.g. 100 views or runs) is split
into multiple sub-counters (e.g. 10 x 10 views).
A cryptographic protocol which is suitable for
secure and duplication-proof transfer of license from
A to B is described in (Carlsen, Hammerstad &
Gorancic, 2003). This protocol has the properties
that the licensed is transferred no more than once to
the rightful destination token, and may never be
transferred to any other token.
The online backup server could be hosted by the
vendor / content owner. It could also be a smart card
which could then be located in the home domain
indicated in figure 4. The sub-licenses are
transferred to and stored on the mobile device SIM
card. Now, if the mobile device SIM card is lost or
corrupt the damage is reduced since only sub-
licenses are stored here. Furthermore, the time
interval
Δ
t can be made small to minimise the
consequences of a lost or corrupted SIM card.
4.4 Requirements for Future SIM
Cards
The paper has demonstrated content rights
management as a potential application for future
SIM cards. SIM functionality is not limited to DRM,
but can have a more general functionality, being the
security infrastructure of the user in the digital world
(Lopez Calvet, 2005). The basic requirement for
such functionality is that the SIM can act on behalf
of the user, supporting different roles and
functionalities. This includes 3
rd
party security
applications, e.g. keys for admittance to buildings,
authorisation to banks and other legal entities.
We suggest using a hierarchical structure, where
the SIM has a master identity module (see figure 5),
which controls the secrets.
Such an identity module might be administrated
by a legal entity, e.g. a network operator, a bank, or
the state. It can then allow administrate secret keys
for other applications, e.g. update, backup, restore.
We see this functionality happen over a secure
channel over the air, and indicated the functionality
by over the air application (OAA).
Admittance and license management are two
items which will require clock (CLK) functionality,
LICENSE TRANSFER MECHANISMS THROUGH SEAMLESS SIM AUTHENTICATION - Rights Management based
on SIM Keys
337