Analysis and Design of Fault-Tolerant Computer Systems
Abstract
This report describes the first-year results of an investigation of fault-tolerant computer systems. A new method for measuring recovery time in fault-tolerant multiprocessors was developed. A complete characterization of optimally t-step recoverable systems was obtained, and certain graph transformations that simplify recovery analysis were studied. Some diagnosability properties of n-cube interconnection networks were derived. A study of fault tolerance in large connecting networks was initiated using a new concept of dynamic full access. A design theory based on recursive component expansion capabilities was developed for MSI/LSI systems. The use of similar recursive methods for test pattern generation was also initiated. Promising results were obtained for testing bit-sliced microprocessors and related components.
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 01, 1978
- Accession Number
- ADA059936
Entities
People
- John P. Hayes
Organizations
- University of Southern California