Databases Reference
In-Depth Information
The model proposed in [26] does not use SQL syntax, but exploits the
concept of view as in relational databases to restrict access to data. In this
case, views are defined by using the XQuery language, and may be authoriza-
tion objects. The model supports not only structure-based authorizations, but
also rules depending on the context or content of the considered documents
by adding conditions in XQuery expressions.
Since relationship among elements/attributes may reveal sensitive infor-
mation, in [27] the authors propose the definition of access control rules on the
relationship among XML elements and attributes (i.e., on arcs in the XML
tree). It is then presented a technique to control the view that can be released
of the
path
leading to any authorized node in an XML document. The authors
introduce also a rule-based formulation of the new class of authorizations.
To the aim of adding semantic meaning to authorizations, RDF (Resource
Description Framework) is used as a new way for expressing access control
policies [28]. The paper focuses also on the problem of controlling
data associ-
ations
, and adds a new object type to the classical model: the association secu-
rity object. An association security object is an XML subtree whose elements
can be accessed only separately. To solve the problem of data associations,
the model uses temporal data.
All the models introduced above for access control of XML documents
are based on the discretionary access control model [12]. In [29], the authors
propose a role-based access control model (RBAC) for XML documents, which
exploits the main characteristics of XML data.
In [30] the authors propose the first access control model for XML docu-
ments operating client-side. The main difference with respect to the previous
proposals is that this method needs to operate on stream data and it is sup-
posed to operate on a system where the server storing data may not be trusted
for access control enforcement.
Recently, a new class of methods have been also proposed for access con-
trol enforcement for XML documents [5, 31, 32]. These methods consider a
data outsourcing scenario, where XML documents are stored on a possibly
not trusted server, and are not under the data owner's direct control. In these
cases, XML documents themselves should enforce access control, since this
task cannot either be executed by the owner or by the storing server. Access
control is enforced through selective encryption, that is, by encrypting differ-
ent portions of the XML tree by using different encryption keys. Consequently,
a correct key distribution to users ensures that access control enforcement is
correct.
5 Conclusions
The role of XML in the representation and processing of information in cur-
rent information systems is already significant and is certainly going to see
a considerable increase in the next years. The design and implementation of
