Design of fault-tolerant protocols for distributed processing systems
This thesis presents a unified approach to the problem of fault tolerance. The author surveys the methods used in the design of distributed algorithms for solving problems in fault-free networks and observes that computing a tree function of the private values of the processors is common to a large class of algorithms. He makes this paradigm fault-tolerant and consequently converts the existing algorithms into fault-tolerant algorithms. Link failures are considered first. It is shown that it is possible to deal with link failures even when the network is asynchronous. The techniques is then applied to a sample set of problems and fault-tolerant protocols are obtained. The problems considered include system problems, set theoretic problems, and topological problems. For one of the topological problems, namely, constructing shortest paths in fault-free networks, several issues are considered and tight lower and upper bounds are presented. Finally, the case of node failures is considered.
- Research Organization:
- Pittsburgh Univ., PA (USA)
- OSTI ID:
- 6988027
- Country of Publication:
- United States
- Language:
- English
Similar Records
A fault-tolerant mapping scheme for a configurable multiprocessor system
Fault tolerance and reliability analysis of large-scale multicomputer systems
Related Subjects
990200* -- Mathematics & Computers
ALGORITHMS
COMPUTER NETWORKS
COMPUTERS
DATA PROCESSING
DIGITAL COMPUTERS
DISTRIBUTED DATA PROCESSING
FAILURE MODE ANALYSIS
FAULT TOLERANT COMPUTERS
MAPPING
MATHEMATICAL LOGIC
PROCESSING
SYSTEM FAILURE ANALYSIS
SYSTEMS ANALYSIS
TOPOLOGICAL MAPPING
TRANSFORMATIONS