In Partial Fulfillment of the Requirements for the Degree of
Doctor of Philosophy
Will defend his pre-defense
Real-time systems refer to the domain of study in computer science where the completion time of any computational activity is placed under certain constraints. The primary constraint is the amount of time taken for the activity to complete, while other constraints can be CPU, power and other system resources used in the processing of that activity. For a real-time system, if any process is unable to complete by a certain time, it is assumed to have not executed at all. All mission-critical systems fall in the domain of real-time systems, examples of which are found everywhere, from anti-locking brakes in cars to fly-by-wire systems in modern aircraft. Many embedded systems like pace makers and sensors also have real-time attributes. Functional Reactive Programming (FRP) is a declarative approach to modeling and building reactive systems. FRP has been shown to be an expressive formalism for building graphics, music, robotic, and vision applications. Recently, Priority-based FRP (P-FRP) was introduced as a formalism that guarantees real-time response. To date, however, no accurate analysis on response time and energy use has been done, without which the practical use of P-FRP is limited. Our work provides a comprehensive response time and power analysis for P-FRP. For the response time analysis part we will develop theory for scheduling conditions, response time bounds, determining exact response time and feasible priority assignment of events in P-FRP. For systems with multiple processors we propose to develop scheduling policies that are optimized for the P-FRP execution paradigm. We will also analyze the power requirements of P-FRP and develop algorithms that implement dynamic voltage scaling for this execution paradigm.