Towards Access-Control Policies Written By Managers
Michael Huth ( Imperial College, London )
- 16:30 21st October 2008 ( week 2, Michaelmas Term 2008 )Lecture Theatre B
Today most access-control policies reside within IT systems - rule files of firewalls being an example thereof. Thus policy writers are mostly technically skilled people. Tomorrow more policies will
have to be formulated by managers or owners of assets that are not directly represented or identifiable within IT systems - legal audit requirements and obligations in work flows being two examples.
This creates a need for policy languages that can be used and analyzed by people with little or no exposure to information technology.
We design a policy language that goes some way towards this strategic goal. Our language has a small core, policy combinators are methods, policy analysis is based on push-button technology such as SAT solvers or BDDs, and propositional constraints express sets of access requests. Key to the simplicity of this approach is the ability to promote a constraint to a policy (those requests that the policy grants,
respectively, denies), and to demote a policy to a constraint. The foundations of this work are rooted in Belnap's four-valued logic.
This is joint work with Glenn Bruns from Bell Labs, Naperville, Illinois.
have to be formulated by managers or owners of assets that are not directly represented or identifiable within IT systems - legal audit requirements and obligations in work flows being two examples.
This creates a need for policy languages that can be used and analyzed by people with little or no exposure to information technology.
We design a policy language that goes some way towards this strategic goal. Our language has a small core, policy combinators are methods, policy analysis is based on push-button technology such as SAT solvers or BDDs, and propositional constraints express sets of access requests. Key to the simplicity of this approach is the ability to promote a constraint to a policy (those requests that the policy grants,
respectively, denies), and to demote a policy to a constraint. The foundations of this work are rooted in Belnap's four-valued logic.
This is joint work with Glenn Bruns from Bell Labs, Naperville, Illinois.
Share this:
