Security Policy Modelling in the Mobile Agent System

The mobile agent security problem limits the use of mobile agent technology and hinders its extensibility and application because the constantly progressed complexity and extension at the level of systems and applications level increase the difficulty to implement a common security system as well as an anticipated security policy. Ontology is considered one of the most important solutions to the problem of heterogeneity. In this context, our work consists of constructing mobile agent domain security ontology (MASO) in order to eliminate semantic differences between security policies in this domain. We use the OWL language under the protected software to construct this ontology. Then, we chose the WS-Policy standard to model security policies, these policies are structured in forms of security requirements and capabilities. To determine the level of semantic correspondence between security policies we are developing an algorithm called "Matching-algorithm" with Java language and two APIs (Jena API and Jdom API) to manipulate the MASO ontology and security policies.