Abstract
Two major steps in the evolution of autonomous, long-life spacecraft computing systems have been the demonstrations of systems with one-year and ten-year life expectancies in space. This paper discusses the design concept of a distributed, diversified self-testing and self-repairing computing system that is embedded in an autonomous spacecraft and serves as its agent for automatic maintenance. The attainment of a hundred-year life expectancy is founded on the extensive use of diversity in materials, technologies and design of hardware elements. Diversity of hardware and software designs also provides tolerance of design faults. A hierarchical system structure and a rigorous design paradigm for the fault-tolerance defenses also support the goals of autonomy and long system lifetime.