CCE Theses and Dissertations


Computer-Aided Performance Analysis using Product-Form Queueing Networks to Model Steady-State Behavior: An Examination of a Medical Device Communications Network

Date of Award


Document Type


Degree Name

Doctor of Philosophy (PhD)


Graduate School of Computer and Information Sciences


Marlyn Kemper Littman

Committee Member

Sumitra Mukherjee

Committee Member

Gregory Simco


Healthcare information systems share records through common messaging standards and exchange information via universal network communications protocols. This interaction benefits hospitals by lowering administration costs and improving the accuracy of recorded information. This sharing and exchanging of information benefits patients by providing easier access to medical records, enabling point-of-care services, and simplifying retrieval of real-time patient data resulting in better patient care. Bedside medical device data complements the overall healthcare information system by providing a more complete understanding of a patient's health. This dissertation presents a simulation based on a standard network communications protocol for medical devices. This researcher addressed the problem that the use of the IEEE 1073 communications protocol to facilitate communication between legacy medical devices and hospital information systems is not adequately understood. The goal of this research was to further develop the understanding of such a system. The objective of this study was to develop a simulation model based on a reasonable approximation of a hypothetical system to identify the parameters involved and quantify its performance characteristics using real-world inputs.

A model was developed using product-form queueing networks to model the steady-state behavior of a medical device communications system. The simulation model consisted of elements representing both physical and logical resources. Only the Physical and Data-Link Open Systems Interconnection (OSI) layers were considered. System configurations were limited to those defined by the IEEE 1073 communications protocol that support legacy medical devices. The inputs to the model consisted of real –world information compiled from vendor data specifications including physical communication mediums, microprocessors, medical devices, and representative software implementations. The results of the simulation suggest that a medical device communications network employing an IEEE 1073 communications protocol can support a limited number of legacy medical devices assuming a I -limited round-robin scheduling policy with a store and- forward data coherency strategy. However, under heavily loaded conditions, it is inadequate to deliver data generated from periodic multi -class work-loads in a timely manner. This conclusion demonstrates the need for an efficient medium access protocol schedule specification.

This document is currently not available here.

  Link to NovaCat