Home

Courses

Publications

Research

CV (pdf)

Contact Info



Research Interests

My current research interests lie in reasoning about access control policies on the web. I am interested in the trade-off between expressiveness of the policy language and the analysis services that can be provided for it. Currently, I am exploring using decidable subsets of First-Order logic (Description Logics or Datalog) to provide analysis services for expressive web services policy languages (such as WS-Policy and XACML). More information is available in my thesis proposal below. Also, you can visit our WS-Policy project page or visit my publications page for more info.


Thesis Proposal

With the widespread use of web services, there is a need for adequate security and privacy support to protect the sensitive information these services could provide. As a result, there has been a great interest in access control policy languages for web services which accommodate large, open, distributed and heterogeneous environments like the web. Currently, the languages with largest momentum include XACML (an expressive, general purpose XML-based language that can be used to specify policies on web resources) and WS-Policy, a policy language for web services.

With these expressive languages, a new problem has emerged: it becomes much harder for security administrators to understand the overall effects and consequences of their policies. Even checking that the access policy will not result in the leakage of permissions to an unintended or unauthorized principal ( i.e., safety enforcement ), has become difficult, if not impossible to do manually. Underspecified security policies might unintentionally give access to an intruder. How can a security administrator be certain that her policy covers all possible corner cases? In the cases when the administrator does discover a bug and fixes the policy accordingly, the consequences of that fix (change) are difficult to analyze. It is possible that due to subtle interactions of some other policy rules with the fix, other security holes are opened. To be certain their security policies express the desired safety requirements, policy administrators could benefit from automated tools that can formally verify policies against safety properties and suggest possible fixes in cases errors are found.

The goal of my research is to develop a logic-based framework that provides means for automated analysis and verification of expressive access control policy languages for web services. As basis for this framework I use Description Logics (DL), which are a family of formalisms that are decidable subsets of First-Order logic, and the basis for theWeb Ontology Language (OWL). Because of the correspondence of policy analysis services to DL reasoning services (e.g., change analysis and verification can be reduced to concept satisfiability and policy containment to subsumption), the framework can provide a variety of novel services and leverage off-the-shelf optimized DL reasoners. Being based on Description Logics, the policy framework is a) expressive, because it captures common policy idioms such as separation of duty, role cardinality constraints, role hierarchies, conflict resolution and negative authorizations, and b) analyzable because DL can provide all of the aforementioned services in a manner more efficient than using full FOL theorem provers. The analysis framework also benefits from the web-like nature of OWL-DL, which makes it convenient for expressing policies for web services.