Future Work
Exploring the limits of what policies this framework supports.
Can we optimize the overhead of state-dependent decisions? Essentially,can we emulate the semantics possible with a centralized controller at an acceptable cost?
What control functions need to be made available to security and policy modules?