Online Model-Based Clustering for Crisis Identification in Distributed Computing

Large-scale distributed computing systems can suffer from occasional severe violation of performance goals; due to the complexity of these systems, manual diagnosis of the cause of the crisis is too slow to inform interventions taken during the crisis. Rapid automatic recognition of the recurrence o...

Full description

Saved in:
Bibliographic Details
Published in:Journal of the American Statistical Association 2011-03, Vol.106 (493), p.49-60
Main Authors: Woodard, Dawn B., Goldszmidt, Moises
Format: Article
Language:English
Subjects:
Applications
Applications and Case Studies
Approximation
Bayesian analysis
Bayesian inference
Bayesian method
Computational methods
Computers
Data analysis
Dirichlet problem
Distributed computing
Distributed processing
Email
Estimate reliability
Exact sciences and technology
General topics
Inference
Inference from stochastic processes
time series analysis
Information systems
Information technology
Internet security
Markov chains
Mathematics
Matrices
Modeling
Monte Carlo simulation
Multivariate analysis
Probability and statistics
Sciences and techniques of general use
Statistics
Time series
Time series models
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Large-scale distributed computing systems can suffer from occasional severe violation of performance goals; due to the complexity of these systems, manual diagnosis of the cause of the crisis is too slow to inform interventions taken during the crisis. Rapid automatic recognition of the recurrence of a problem can lead to cause diagnosis and informed intervention. We frame this as an online clustering problem, where the labels (causes) of some of the previous crises may be known. We give a fast and accurate solution using model-based clustering based on a Dirichlet process mixture; the evolution of each crisis is modeled as a multivariate time series. In the periods between crises we perform full Bayesian inference for the past crises, and as a new crisis occurs we apply fast approximate Bayesian updating. These inferences allow real-time expected-cost-minimizing decision making that fully accounts for uncertainty in the crisis labels and other parameters. We apply and validate our methods using simulated data and data from a production computing center with hundreds of servers running a 24/7 email-related application.
ISSN:0162-1459
1537-274X
DOI:10.1198/jasa.2010.ap09545