Off-campus UMass Amherst users: To download dissertations, please use the following link to log into our proxy server with your UMass Amherst user name and password.
Non-UMass Amherst users, please click the view more button below to purchase a copy of this dissertation from Proquest.
(Some titles may also be available free of charge in our Open Access Dissertation Collection, so please check there first.)
Fault diagnosis and masking in distributed computing systems
Abstract
We consider issues of fault tolerance for distributed computing systems at two levels of system design: (1) system integration and (2) distributed software. Corresponding advances are made in: (1) reducing testing overhead when employing system-level diagnosis under the PMC model and (2) handling multiple types of failures in distributed agreement protocols. We consider system-level diagnosis for distributed systems. We treat fault diagnosis as two distinct processes: fault discovery and dissemination of diagnostic information. Testing overhead consists of periodic tests for fault discovery and further tests and message passing for dissemination. We discuss homogeneous and nonhomogeneous systems in test-synchronous and asynchronous environments. We present a new methodology that minimizes the periodic testing overhead. Minimizing periodic testing allows less testing overhead, greater test reliability, or more frequent testing. The method is applied to a variety of systems, including nonhomogeneous systems and weighted systems. We discuss interactive consistency for distributed systems. Systems are treated where either or both of two types of faults may occur: benign (very well controlled faults) and malicious (may exhibit arbitrary behavior). We establish a continuum between previous results that allowed either benign or malicious faults to occur separately. We give an algorithm and compare it to known algorithms. We give a method to establish virtual links for graphs not completely connected. Virtual links are a limiting factor in achieving consensus. We introduce a family of graphs with the widest containers amongst graphs in the literature. We compare the graphs with respect to: diameter, fault-tolerant diameter, container width, ease of routing, ease of fault-tolerant routing, and extensibility.
Subject Area
Computer science
Recommended Citation
Meyer, Fred John, "Fault diagnosis and masking in distributed computing systems" (1991). Doctoral Dissertations Available from Proquest. AAI9207438.
https://scholarworks.umass.edu/dissertations/AAI9207438