Formal Techniques for Fault-Tolerance in Distributed Data Processing (DDP).
Jack H. Goldberg, William H. Kautz, Leslie Lamport, Peter G. Neumann, Robert E. Shostak · Defense Technical Information Center (DTIC) · 1979
Distributed data processing can lead to systems with greatly improved reliability when compared with conventional systems, and can offer flexible strategies for redundancy, self-testing, and reconfiguration. DDP systems also have potential advantages in security, efficiency, and evolvability. The work in this contract is aimed at providing a suitable theoretical basis for the development of practical techniques for DDP system design and verification. Three fundamental issues have been investigated and are reported here: hierarchical structure in DDP systems; achievement of consistent redundant data representations in faulty networks--specification, design, and validation; and self-diagnosability in DDP networks.