The number and complexity of embedded systems is rapidly increasing. Especially for large distributed real-time systems, the determination of a feasible deployment is a complex task. In this thesis we present a design approach for distributed real-time systems that supports the designer to determine an appropriate deployment.Distributed systems with hard real-time constraints are so-called safety-critical systems. In safety-critical systems, missing a hard deadline may cause catastrophic consequences. A system must also reliably provide its intended functionality. Fault tolerance enables a system to continue a safe and reliable operation even in presence of a fault. We focus on the compensation of hardware faults during system runtime resulting in network and processor failures. All fault-tolerant techniques require additional redundant elements to detect and compensate faults. Hardware redundancy is perhaps the most common form of redundancy and can be categorized into static and dynamic redundancy. We apply dynamic redundancy by means of reconfiguration to realize fault tolerance. We present concepts for a reconfigurable network topology and for the coordination of the required reconfigurations. Based on these concepts, we extend our approach to design fault-tolerant distributed real-time systems. We apply our approach to the de facto standard AUTOSAR for automotive system design and approve its feasibility for realistic systems by means of a real-world case study.