Title :
Perturbation analysis for optimal update intervals of data sets
Author :
Ruschmann, Matthew C. ; Wu, N. Eva
Author_Institution :
Dept. of Electr. & Comput. Eng., Binghamton Univ., Binghamton, NY
Abstract :
Update intervals for data renewal in a 3-server distributed database are optimized to provide maximum system availability. A semi-Markov model of the database unit is defined. Given probability distributions, analytical expressions are derived for the expected availability of the database. Under the case of unknown failure rates and renewal intervals, estimates for the gradients of the sample performance functions are derived using infinitesimal perturbation analysis. Optimal values of the update interval are found using a stochastic approximation algorithm.
Keywords :
Markov processes; approximation theory; distributed databases; set theory; statistical distributions; 3-server distributed database; data sets; maximum system availability; optimal update intervals; perturbation analysis; probability distributions; semi-Markov model; stochastic approximation algorithm; Approximation algorithms; Availability; Control systems; Distributed databases; Failure analysis; Helium; Optimal control; Performance analysis; Probability distribution; Stochastic processes;
Conference_Titel :
American Control Conference, 2008
Conference_Location :
Seattle, WA
Print_ISBN :
978-1-4244-2078-0
Electronic_ISBN :
0743-1619
DOI :
10.1109/ACC.2008.4587173
