1st International ICST Workshop on Tools for solving Structured Markov Chains

Research Article

A Tool for the Analysis of Hierarchical Service-Oriented Extended Open Fork/Join Queueing Networks

Download682 downloads
  • @INPROCEEDINGS{10.4108/smctools.2007.1910,
        author={Markus Arns},
        title={A Tool for the Analysis of Hierarchical Service-Oriented Extended Open Fork/Join Queueing Networks},
        proceedings={1st International ICST Workshop on Tools for solving Structured Markov Chains},
        proceedings_a={SMCTOOLS},
        year={2010},
        month={5},
        keywords={Fork/Join Queueing Networks QBD-Analysis Aggregation},
        doi={10.4108/smctools.2007.1910}
    }
    
  • Markus Arns
    Year: 2010
    A Tool for the Analysis of Hierarchical Service-Oriented Extended Open Fork/Join Queueing Networks
    SMCTOOLS
    ICST
    DOI: 10.4108/smctools.2007.1910
Markus Arns1,*
  • 1: Department of Computer Science University of Dortmund D-44221 Dortmund
*Contact email: markus.arms@uni-dortmund.de

Abstract

In this paper we present a tool for the analysis of hierarchical service-oriented extended open fork/join queueing networks (EOFJQNs). The tool especially focusses on application areas that consider a service-oriented respectively processoriented view of systems. Such application areas are for example business process management, logistics, supply chain management, production planing and control, and computer and communication systems. Since parallel process execution is a typical property of corresponding models we consider extended queueing networks that incorporate complex fork/join structures. For the analysis of EOFJQNs we apply a frequently used decomposition approach and we additionally approximate inter-arrival times and service times with phase-type distributions. Then the analysis of the isolated nodes is based on their underlying quasi-birth-and-death process. In this context the main problem is the analysis of complex fork/join structures. Therefore, at first we briefly reflect an approach for the analysis of rather simple fork/join nodes and afterwards apply a new aggregation techniques that allows us to reduce complex fork/join structures to simple fork/join nodes. We apply our tool to the analysis of a parallel production line.