Publication Date


Degree Type

Master's Project

Degree Name

Master of Science (MS)


Computer Science


While it takes traditional internet worms hours to infect all the vulnerable hosts on the Internet, a flash worm takes seconds. Because of the rapid rate with which flash worms spread, the existing worm defense mechanisms cannot respond fast enough to detect and stop the flash worm infections. In this project, we propose a geometric-based detection mechanism that can detect the spread of flash worms in a short period of time. We tested the mechanism on various simulated flash worm traffics consisting of more than 10,000 nodes. In addition to testing on flash worm traffics, we also tested the mechanism on non-flash worm traffics to see if our detection mechanism produces false alarms. In order to efficiently analyze bulks of various network traffics, we implemented an application that can be used to convert the network traffic data into graphical notations. Using the application, the analysis can be done graphically as it displays the large amount of network relationships as tree structures.