An Algorithm for Classifying PSTN Switching Stations.
Abstract
The National Security Agency (NSA) collects and processes signals intelligence information for national security purposes. As part of this mission, NSA predicts message routing over public switched telephone networks (PSTNs). The hierarchical switching level (or classification) of PSTN switching stations must be determined before making routing predictions. This thesis develops a fast graph-theoretic algorithm for accomplishing this classification. An undirected connected graph models a target PSTN; switching stations are nodes and logical connections between the switching stations are unit-length arcs. We develop bounds for the minimum number of switching levels and implicitly enumerate all possible classifications for each PS TN. The algorithm is implemented in Java and PSTNs are classified using a personal computer. Solutions are obtained in under one second for nine real-world PSTNs, and large notional networks of over 300 nodes and 900 arcs are classified in under one minute. This research improves existing node classification software.
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 01, 1998
- Accession Number
- ADA355957
Entities
People
- John F. Brandeau
Organizations
- Naval Postgraduate School