ABSTRACT
This chapter addresses two important problems. The first is the need for Digital Signal Processing (DSP) systems with higher levels of both performance and fault tolerance. The second is the need for methods to design and develop new multiprocessor and distributed systems in a way that reduces the escalating costs of this extremely complex task. In addressing these two problems, three objectives were identified. The first objective was the creation of a new system for DSP which provides higher levels of performance and fault tolerance over traditional single-processor or non-redundant DSP designs. The second objective was the implementation of a new system prototype without the prohibitive costs normally associated with such designs. The third objective was to evaluate DSP new system with respect to both performance and fault tolerance. In addition to improvements in DSP performance, reliability modeling of this new system has shown a substantial improvement in reliability and mission time for this system versus a comparable non-redundant system.
