In Chapter 1, a brief history of REL is given. A deeper dive into major branches of RELs helps to understand the challenge of creating a generic rights model for rights exporting. As ODRL and XrML seem to be the foundation of two major branches in REL (Schmidt et al. 2004), this section mainly reviews the essential features of those RELs and their related research.
XrML is based on DPRL developed by a group at Xerox Palo Alto research center (Guo 2001). In 1998, Xerox released the XML implementation of DRPL as Version 2.0 (DPRL 1998). In 2000, ContentGuard, a joint venture between Xerox and Microsoft, released XrML Version 1.0 as an evolution of DRPL (Jamkhedkar et al. 2006). XrML Version 2.0 was selected as the basis for REL standard development of MPEG21 Part 5 as well as the REL standard for the Open eBook Forum (Guth et al. 2003). In 2002 XrML Version 2.1 was submitted to OASIS with the intention to serve as the basis for defining the industry standard REL (Schmidt et al. 2004).
ContentGuard has discontinued further development of XrML and transferred the responsibility to standardization bodies like OASIS Rights Committee and MPEG initiative (Guth et al. 2003).
XrML is an XML-based declarative language for expressing rights granted by some principals for specific resources and the conditions under which those rights apply (Wang et al. 2002). There are four main entities defined in XrML data model:
Principal is the party to which rights are issued to.
Resource is the content of the rights.
Right specifies what is granted to the principal.
Condition must be met before rights can be consumed.
XrML basic data construct is a grant. A grant consists of the instances of the four main entities. A license holds a set of grants with license issuer’s info as well as other license specific data as illustrated in Figure 19.
Figure 19. XrML rights expression model (Wang et al. 2002)
The syntax of XrML 2.0 is described and defined with XML Schema technology (W3C 2001), which enables XrML to provide a high degree of richness and flexibility in its expressiveness and extensibility (Wang et al. 2002). The defined schemas in XrML consists of the following:
Core schema contains core definitions, particularly related to the evaluation of a trust decision.
Standard extension schema contains definitions applicable to usage scenarios
Content extension schema contains definitions of rights management related to content.
Those defined XML schemas are normative parts of the XrML specification. By using the existing standard XML schema and XML namespace, other parties may define their own extensions to XrML 2.0. The existing and future extensions are illustrated in Figure 20.
Figure 20. Schemas in XrML (Wang et al. 2002)
ODRL Version 0.5, an XML-based REL was proposed by Renato Iannella (Iannella 2004) from IPR System Ltd in 2000 (Guth et al. 2003). In 2001 ODRL Version 1.0 was released in response to a call for rights data dictionary (RDD)-REL (Jamkhedkar et al. 2006). ODRL was influenced by RealNetworks’ eXtensible Media Commerce Language (XMCL) and Nokia’s Media Rights Voucher (MRV) (Schmidt et al. 2004).
In 2002 ODRL Version 1.1 was adopted by Open Mobile Alliance (OMA) as the REL used in its DRM specifications (OMA DRM Version 1.0) for mobile content (Iannella 2004).
ODRL is based on an extensible model for rights expressions and includes the following core entities (Iannella 2004):
Assets include physical or digital content that can be uniquely identified.
Rights include permissions.
Permission is the actual usage or activities granted over assets. Permission contains constraints, requirements, and conditions.
Constraints are limitations to Permissions
Requirements are obligations needed to exercise permissions
Conditions specify the exceptions that expire the permissions once they turned true.
Parties include end users and rights holders.
End users are asset consumers
Rights holders are parties that have played some role in the creation, production, distribution of the Asset.
Figure 21. ODRL rights expression model (Iannella 2004)
ODRL specification declares two XML namespace for two XML Schemas:
o-ex for expression language schema o-dd for data dictionary schema
The data dictionary is a collection of terms for core entities. Through data dictionary, the semantics can be extended or reused. For example, Iannella (2004) demonstrated how to add extensions to the standard ODRL data dictionary.
As both XrML and ODRL are XML based, researchers have proposed a solution to translate between these two RELs. Polo et al. (2004) proposed a solution by directly transforming XML documents to translate licenses between ODRL and XrML. However, they claimed that the solution was able to handle only simple licenses. Cooper and Montague (2005) did a similar study on translating between XrML and ODRL. They highlighted the challenges in three categories:
Cooper and Montague (2005) also argued on the ambiguities in ODRL definition and proposed to either to translate from a profile of ODRL to XrML e.g. from the OMA REL profile in order to limit the elements contained in the source REL, or to translate to an intermediate schema to mimic the structure of source REL while utilizing the core entities from the target REL.
Jamkhedkar et al. (2006) argued that in order to support interoperability the rights expression functionality should be as simple as possible. They believed that current RELs like XrML and ODRL had been developed over time to facilitate a set of DRM services, such as authentication and payment mechanism, besides the rights expression as the core service that RELs need to serve. The data required for those services and management mechanism are combined as illustrated in Figure 22. Thus, the semantics required for those services are merged into the RELs. It makes RELs extremely difficult to change those services without changing themselves. In other words, it makes it difficult to interoperate between two DRM systems based on the different RELs.
Figure 22. DRM functionalities implemented using existing RELs (Jamkhedkar et al. 2006)
Jamkhedkar et al. (2006) concluded their recommendation on the design principles related to the core REL:
The core REL should only contain the rights model.
The core REL should be stateless.
The core REL should be language neutral.
The core REL should refer generically to the protocol or service they use instead of being specific.
A core REL provides a separation of rights management mechanisms from the rights expression, which is essential for the encapsulated DRM services that will be necessary to achieve DRM interoperability. As Jamkhedkar et al. (2006) suggested, a core REL should only contain the rights model. A generic rights model is, therefore, the essential part of DRM interoperability. On the other hand, this dissertation wants to discuss the nature of rights in a REL-agnostic manner so that the research result can be applied to systems based on different RELs. Thus, a generic rights model needs to be established.