ABSTRACT
To our knowledge, no prior work deals with multilevel semantic Webs. Much of the work on multilevel information systems has focused on multilevel data-management systems where a user is cleared at different clearance levels and the data is assigned different sensitivity levels. Users read data at or below the level and write data at their level. It is assumed that these sets of levels form a particularly ordered lattice, with Unclassified < Confidential < Secret < Top Secret. In this chapter we will discuss multilevel security issues for the semantic Web. We will extend the work on multilevel secure data-management systems to address multilevel security for XML, RDF, and ontologies as well as reason about these different technologies so that security violations via inference does not occur.
