| Single-resource and distributed schedulability analyses verify that a real-time system will meet its timing requirements, whether they be individual task deadlines or end-to-end latency requirements. This thesis presents a formal modeling framework that specifies the necessary information for describing both the service time requirements of tasks on a single resource and distributed computing systems. The framework is aligned with software engineering design processes, and provides mechanisms for the software architect to readily conduct schedulability analysis.; The system model contains a set of application models and a target-platform model. Both the application models and target-platform model are broken into domain, architecture, component, and physical levels. The domain level describes key design parameters affecting system performance. The architecture level captures the divide-and-conquer design strategy hierarchically. The component level describes individual target- platform resources or single tasks in the application. The physical level relates the application and target- platform models and reflects the physical world. Individual target-platform components are modeled as a combination of hardware, system software, and concurrency parts, as this work views hardware, system services, and application software as separable entities. The application software component model, which describes the service time requirements of a single schedulable entity, uses both the hardware and system software services of a target-platform component. The application software component model captures four types of service time: constant, scenario driven, cyclic, and stochastic. The modeling framework provides data for multiple distributed system analysis and single resource analysis algorithms.; Experiments on a real-time operating system with multimedia applications verify the abilities of the application software component model to capture service times accurately and reliably and serve as a source for diverse schedulability analysis techniques. |
