ABSTRACT:

The universe model is a new approach to managing concurrency that
isolates concurrency concerns and represents them as declarative
statements in the modular interface of a component. This approach
facilitates program comprehension, composability, and reliability for
concurrent systems. It is founded on designer-specified invariant
properties that declare a component's dependencies on other concurrent
components.

Process scheduling requirements can be derived automatically from these
properties, but scheduling remains a non-trivial problem because of
fairness considerations. We will discuss how process scheduling in the
universe model can be treated as a resource allocation problem, and how
a classic textbook algorithm can be adapted and extended to guarantee
fairness properties in the more general case of the universe model.