A probabilistic approach to measuring robustness in distributed computing systems
Keywords and Phrases
Robustness metric; Mahalanobis distance
"The robust design of computing and communication systems is becoming an increasingly important issue. Parallel and distributed heterogeneous computing systems may operate in dynamic environments that undergo unpredictable changes causing certain system performance features to degrade. Such systems need robustness to guarantee limited degradation despite fluctuations in behavior of its component parts or environment. There is a need for research that addresses the issue of developing a generalized robustness metric. This thesis describes a probabilistic approach to measuring robustness such that variance and correlation are considered in the robustness measure. In addition, this research presents a methodology where perturbations of different kinds, e.g., in both message sizes and computation times, could be incorporated in the robustness metric. This will be an important step towards ongoing efforts to create robust designs"--Abstract, leaf iii.
Electrical and Computer Engineering
M.S. in Computer Engineering
University of Missouri--Rolla
Journal article titles appearing in thesis/dissertation
- Incorporating multiple kinds of perturbations in the robustness measure
- Incorporating probabilistic information in the robustness measure
viii, 44 leaves
© 2005 Behdis Eslamnour, All rights reserved.
Thesis - Citation
Library of Congress Subject Headings
Computer systems -- Reliability
Electronic data processing -- Distributed processing
Parallel processing (Electronic computers)
Print OCLC #
Link to Catalog Record
Eslamnour, Behdis, "A probabilistic approach to measuring robustness in distributed computing systems" (2005). Masters Theses. 5822.