Date of Award


Document Type


Degree Name

Master of Science


Department of Electrical and Computer Engineering

First Advisor

Kennard R. Laviers, PhD.


With increasing magnitude of computer network activity, the ability to monitor all network traffic is becoming strained. The need to represent large amounts of data in smaller forms is essential to continued growth of network monitoring tools and network administrators' capabilities. Network monitoring captures many different measurements of the data flowing through the network. This thesis introduces a new method of sending network traffic monitoring data that reduces the overall volume of data from the traditional method of packet capture. By populating a matrix with specific data values in a sparse format, this experiment reduces the data using singular value decomposition (SVD) compression. Matrices were populated using network monitoring datasets from 1996 Information Exploration Shootout (IES). The data populated into the matrices was varied along time frame and data field to determine if the SVD compression algorithm reduced the quantity of original data values. Results indicated that the quantity of data varies dependent on the volume of the data field chosen. The matrix population method was based on port values to allow combining values within the matrix cells. The results trended to a successful reduction of data if the time frame is increased significantly.

AFIT Designator


DTIC Accession Number