All ETDs from UAB

Advisory Committee Chair

Murat M Tanik

Advisory Committee Members

Gary J Grimes

Laurie L Joiner

Murat Tanju

Gregg L Vaughn

Document Type


Date of Award


Degree Name by School

Doctor of Philosophy (PhD) School of Engineering


Today’s voice and data connectivity typically utilizes a complex system of networks of networks. The development of models for characterizing usage patterns in complex networks would be useful in projecting capacity requirements in growing systems. Simplified models applied to a complex network for timing and sizing algorithms would significantly reduce the amount of computations and storage necessary to produce forecasts of capacity requirements. This composition introduces a network usage pattern model based on decomposition by information transfer and verifies its application in real-life voice-data networks. Complex systems can be modeled and decomposed into sub-systems by observing the interactions among their elements. A normalized transmission parameter is used in this study as the model for comparing sets of measurement data to model instances. Methods for constructing example instances and the method of comparison are described. Measurement data for five voice message trunk groups and ten data circuits is analyzed using three different instances. Validation is accomplished using the model instances to predict the parameters of combinations of traffic usage and comparing the predictions to calculated parameters of the usage combinations. Results of modeling usage data for telephony trunks and internet usage data for two types of circuits are described. Time-consistent busy-hour model instances are compared to 24-hour model instances for each case. For one of the Internet circuit types, iv a third model instance with a 6-hour busy period is included. Time-consistent busy-hour instances had the lowest valued transmission parameters. The 24-hour instances had the highest valued transmission parameters and the 6-hour busy period instances had values in between. Instances with greater transmission parameters yielded more accurate predictions when combinations of measurement data had non-coincident usage patterns. Study results support the original hypothesis that development of models for characterizing usage patterns in complex networks would be useful in projecting capacity requirements in growing systems. The normalized transmission parameter is a useful predictor of relative accuracy of a model in predicting effects of combining usage on trunks or circuits where there was a significant difference in model instance parameters and trunks or circuits had dissimilar busy hours or busy periods.

Included in

Engineering Commons



To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.