Time and reliability analysis of distributed real-time systems based on RTG model

Real-time systems are often distributed either due to the nature of applications or for improving reliability and performance or both. The distributed and concurrent execution of software detracts the predictability of the system behavior. Moreover, in distributed computing environments both reliability and performance (including timeliness) of the system are affected by the architecture of hardware and software components and allocation of the software components to the hardware components. The RTG model provides a basis to analyze both reliability and performance of the system.;Time analysis is performed in two ways: feasibility and schedulability. The feasibility analysis checks if a software design is feasible to satisfy the temporal requirement with a given hardware platform. The minimum hardware requirement is determined based on rate analysis and deadline analysis. The schedulability analysis checks if a feasible application is schedulable. This includes task allocation and scheduling policies for both tasks and messages.;Most studies on reliability analysis have focused on hardware analysis. This research proposes reliability analysis in two different phases of the system development. In the early design phase, the reliability analysis must be an aid to determine the software as well as hardware fault tolerance mechanisms which will be incorporated into the system. In the later phase the implemented system is validated if it satisfies the reliability requirements.;The RTG model along with the RTGTool offer designers a comprehensive paradigm to model and analyze distributed real-time systems. The significance of the approach is the ability (1) to treat both reliability and performance analysis in different phases of the system development, and (2) to consider both software and hardware architecture for the analysis.