Author

Gary K. Moy

Date of Award

12-1996

Document Type

Thesis

Degree Name

Master of Science

Department

Department of Electrical and Computer Engineering

First Advisor

Richard A. Raines, PhD

Abstract

Communications have always been a crucial part of any military operation. As the pace of warfare and the technological complexity of weaponry have increased, so has the need for rapid information to assess battlefield conditions. Message passing across a network of communication nodes allowed commanders to communicate with their forces. It is clear that an accurate prediction of communication usage through a network will provide commanders with useful intelligence of friendly and unfriendly activities. Providing a specific network link and path likelihood prediction tool gives strategic military commanders additional intelligence information and enables them to manage their limited resources more efficiently. In this study, Dijkstra's algorithm has been modified to allow the Queueing Network Analyzer's (QNA) analysis output to act as a node's goodness metric. QNA's calculation of the expected Total Sojourn Time for the completion of queueing and service in a node provides accurate measurement of expected congestion. The modified Dijkstra's algorithm in the Generalized Network Analyzer (GNA) is verified and empirically validated to properly deliver traffic. It appropriately generates the fastest traffic path from a start node to a destination node. This implementation includes notification if input parameters exceed the network's processing capability. GNA's Congestion Control displays notification and informs the user certain network input parameters must be lowered (PTR or BSTR) or where certain nodes must be improved to maintain node stability. With this unstable node identification, users can determine which node needs attention and improvements. Once this instability is removed, a good QoS is achieved and analysis proceeds.

AFIT Designator

AFIT-GCS-ENG-96D-21

DTIC Accession Number

ADA325184

Share

COinS