Automated Rule Tuning 

		Rules are widely used in network management systems. In essence, they allow implementation of the system security policy in a flexible format that can be easily maintained and interpreted by human. However, one of the drawbacks of the rules is the reliance on the user guidance in rule development as well as in their adjustment to the changes in the network environment and security policy.
Current Members:		Natalia Stakhanova
Previous Members:		Shah Arif Iqbal

Overview:

Although the research in the field of rule generation has received a considerable attention, rule adaptation has been left largely unaddressed mainly due to the complexity of the rule refinement in an automatic fashion.

The goal of this project is the automatic and dynamic rule adjustment in response to the changes in the network environment. We focus on two aspects of rule refinement: the adjustment of the rule parameters that involves changes to the values of the conditional part of the rule (e.g. IP addresses, time duration, number of login attempts etc.) and the structural reorganization of the rule that refers to the merge and removal of the existing rules, generalization and specialization of the rule components, etc.