Abstract
Purchase PDF (326 K)
E-mail Article
Add to my Quick Links
Cited By in Scopus (1)
Purchase PDF (2239 K)
doi:10.1016/j.peva.2004.05.004 
Copyright © 2004 Elsevier B.V. All rights reserved.
Configuration of distributed message converter systems
References and further reading may be available for this article. To view references and further reading you must purchase this article.
Thomas Risse
, 
, a, Karl Aberer, b, Andreas Wombacher, a, Mike Surridge, c and Stephen Taylor, c
Received 1 July 2002;
Abstract
Finding a configuration of a distributed system satisfying performance goals is a complex search problem that involves many design parameters, like hardware selection, job distribution and process configuration. Performance models are a powerful tool to analyze potential system configurations, however, their evaluation is expensive, such that only a limited number of possible configurations can be evaluated. In this paper we present a systematic method to find a satisfactory configuration with feasible effort, based on a two-step approach. First, performing a queuing network analysis a hardware configuration is determined and then a software configuration is incrementally optimized by simulating Layered Queuing Network models. We applied this method to the design of performant EDI converter systems in the financial domain, where increasing message volumes need to be handled due to the growing importance of B2B interaction.
Author Keywords: System configuration; Performance modeling; Queuing networks
Article Outline
- 1. Introduction
- 2. Overview of the approach
- 2.1. Application background: message processing
- 2.2. System architecture
- 2.3. System configuration approach
- 2.4. Using exhaustive search
- 3. Performance modeling
- 3.1. Layered queuing networks
- 3.2. Transformation rules
- 3.3. Modeling of a converter system
- 3.4. Simulation tools
- 3.5. Model parameters
- 3.6. Evaluation of the LQN model
- 4. System design method
- 4.1. Goal parameters
- 4.2. Overall approach
- 4.3. Response-time-throughput model
- 4.3.1. Derivation of an upper bound for response time
- 4.3.2. Derivation of a lower bound for response time
- 4.3.3. Derivation of an approximation function
- 4.4. Configuration algorithm
- 4.4.1. Configuration of a single host
- 4.4.2. Configuration of a distributed system
- 4.4.3. Balancing the load distribution
- 4.4.4. Handling of constraints on software configuration
- 4.5. Software configuration
- 4.6. A software configuration algorithm
- 4.7. Implementation
- 5. Application of the configuration method
- 6. Real system measurements
- 6.1. LQN model calibration
- 6.2. Scheduler and configuration tests
- 6.3. Bottleneck analysis
- 6.4. Discussion of experimental results
- 7. Conclusion and outlook
- Acknowledgements
- References
- Vitae
