Bayesian inference for queueing networks and modeling of internet services

@article{Sutton2010BayesianIF,
  title={Bayesian inference for queueing networks and modeling of internet services},
  author={Charles Sutton and Michael I. Jordan},
  journal={arXiv: Machine Learning},
  year={2010}
}
Modern Internet services, such as those at Google, Yahoo!, and Amazon, handle billions of requests per day on clusters of thousands of computers. Because these services operate under strict performance requirements, a statistical understanding of their performance is of great practical interest. Such services are modeled by networks of queues, where each queue models one of the computers in the system. A key challenge is that the data are incomplete, because recording detailed information about… 

Figures and Tables from this paper

A Bayesian Approach to Parameter Inference in Queueing Networks
TLDR
A novel iterative approximation of the normalizing constant of the equilibrium state probabilities is defined and the improved accuracy of this approach is shown, compared to existing methods, for use in conjunction with Gibbs sampling.
Bayesian Service Demand Estimation with Gibbs Sampling
TLDR
A novel approximation is defined to efficiently obtain the normalising constant to make the cost of its evaluation acceptable for MCMC applications, and this expression is shown to be the equilibrium solution of a multiclass closed queueing network.
Bayesian Service Demand Estimation Using Gibbs Sampling
  • Weikun Wang, G. Casale
  • Computer Science
    2013 IEEE 21st International Symposium on Modelling, Analysis and Simulation of Computer and Telecommunication Systems
  • 2013
TLDR
This paper proposes a service demand estimation algorithm based on a Markov Chain Monte Carlo technique, Gibbs sampling, and defines a novel approximation to efficiently obtain the normalising constant to make the cost of its evaluation acceptable for MCMC applications.
Auxiliary variables for Bayesian inference in multi-class queueing networks
TLDR
A slice sampling technique with mappings to the measurable space of task transitions between the service stations is introduced that can address time and tractability issues in computational procedures, handle prior system knowledge and overcome common restrictions on service rates across existing inferential frameworks.
Variational inference for Markovian queueing networks
TLDR
A probabilistic interpretation of variational methods applied to inference tasks with queueing networks is offered, and it is demonstrated that it is still possible to enable a variational inferential task, by considering a novel space expansion treatment over an analogous counting process for job transitions.
Computationally Efficient Simulation of Queues: The R Package queuecomputer
TLDR
This work presents a computationally efficient method for simulating from a very general set of queueing networks with the R package queuecomputer and shows how to use this package with two examples: a call center and an airport terminal.
Maximum Likelihood Estimation of Closed Queueing Network Demands from Queue Length Data
TLDR
This paper proposes new estimators that infer demands from queue-length measurements, which are inexpensive metrics to collect in real systems and are validated against simulation and real system measurements for a multi-tier application.
Approximate Bayesian inference with queueing networks and coupled jump processes
TLDR
This paper offers a comprehensive overview of the inferential challenge and its comparison to analogue tasks within related mathematical domains, and presents a flexible and scalable variational Bayesian framework targeted at general-form open and closed queueing systems, with varied service disciplines and priorities.
Moving Horizon Estimation of Service Demands in Queuing Networks
TLDR
This paper presents the first approach—here developed for single-class QNs—that does not make such assumption and sets up a moving-horizon estimation problem whereby the governing equations of the model, appropriately unfolded over a given time horizon, represent the constraints of a quadratic program that seeks to find the optimal choice of service demands that minimize the error between the measured queue lengths and the predicted ones.
Bayesian analysis of queues with impatient customers: Applications to call centers
TLDR
This article considers Bayesian queuing models with impatient customers with particular emphasis on call center operations and develops the details of Bayesian inference for queues with abandonment such as the M/M/s + M model (Erlang‐A).
...
...

References

SHOWING 1-10 OF 30 REFERENCES
Dealing with Uncertainties in Queues and Networks of Queues: a Bayesian Approach
Queues and networks queues are systems which have been widely and successfully applied to the modelling and study of the performance of complex interacting stochastic systems, in which congestion
An analytical model for multi-tier internet services and its applications
TLDR
This paper presents a model based on a network of queues, where the queues represent different tiers of the application, sufficiently general to capture the behavior of tiers with significantly different performance characteristics and application idiosyncrasies such as session-based workloads, concurrency limits, and caching at intermediate tiers.
Estimating functions in indirect inference
Summary.  There are models for which the evaluation of the likelihood is infeasible in practice. For these models the Metropolis–Hastings acceptance probability cannot be easily computed. This is the
Bayesian prediction inM/M/1 queues
Simple queues with Poisson input and exponential service times are considered to illustrate how well-suited Bayesian methods are used to handle the common inferential aims that appear when dealing
Filtering recursions for calculating likelihoods for queues based on inter-departure time data
TLDR
This work considers inference for queues based on inter-departure time data, and demonstrates how a likelihood recursion can be used to calculate this likelihood efficiently for the specific cases of M/G/1 and Er/G-1 queues.
Network Tomography: Recent Developments
TLDR
This article introduces network tomography, a new field which it is believed will benefit greatly from the wealth of statistical methods and algorithms including the application of pseudo-likelihood methods and tree estimation formulations.
Empirical Bayes estimation for queueing systems and networks
TLDR
Empirical Bayes estimators are derived for standard and finite capacity M/M/1 queues, with state-dependent arrival and service rates, and for open Jackson networks.
Internet tomography
TLDR
This article introduces the new field of network tomography, a field which it is believed will benefit greatly from the wealth of signal processing theory and algorithms.
Inferring Balking Behavior From Transactional Data
TLDR
An extension of Larson's congestion probability calculation to include balking is presented, and new maximum likelihood, nonparametric, and Bayesian methods for inferring the arrival rate and balking functions are presented.
...
...