Integrated Development Environment for Policies

There has been growing interest in the use of policy-based governing frameworks for management of a wide range of systems. These systems ranging from simple and static to increasingly complex and dynamic have demanding requirements that make the management of policies a complex task. Though tools have been developed to facilitate policy frameworks, there is not much work in policy development that meets the requirements of these policy-based environments. Some of these requirements include: (i) facility for both individual and rule-based policy specifications, (ii) ability to test policy conformance, (iii) simplification of the inherently error-prone process of policy creation through the use of templates and well-designed interfaces.

In this thesis, we present RIDE, an integrated policy development environment that uses a wizard-based approach to provide a user-friendly and extensible graphical user interface. RIDE (Rei Integrated Development Environment) has been developed as a plug-in extension of Eclipse workbench and enables the development of policies specified in Rei, which is a declarative, machine-understandable policy specification language. As our main goal was to facilitate policy development for distributed systems, we chose Rei specifications because Rei is grounded in a web ontology language that allows it to model different kinds of domain knowledge and it has been successfully deployed in several open, dynamic distributed environments.

RIDE attempts to meet the requirements of a wide range of policy-based environments for which existing tools provide little or no support. This is possible due to the unique combination of following features it supports: (1) simple, user-friendly interface offering valid input options through the entire process, (2) ease of management of domain information by offering the option of template creation, (3) automation of the complex and error-prone policy creation process by automatically generating user-defined policies in Rei, (4) ability to express individual as well as group policies, (5) ability to accommodate information spanning multiple domains by using ontologies to represent domain information, (6) policy creation over speech acts that are used for dynamic policy management, (7) support for creation and verification of policy test-cases for checking policy conformance, and (8) conformity to the principles outlined by Human Computer Interaction by performing iterative evaluation and refinement of the interface to make it as simple and useful as possible from user's perspective.