A unique military-university partnership is working to improve cybersecurity technology and the safety of military and civilian computer networks. The collaboration includes CUBRC, a not-for-profit research and development company headquartered in Buffalo, N.Y., along with professors from Rochester Institute of Technology, the University of Buffalo and Pennsylvania State University. The research team is seeking to implement the use of intrusion prediction modeling into cyber defense systems.
Intrusion prediction utilizes mathematical models and algorithms to forecast the actions of a computer hacker once they have entered a network. The method provides needed information about how an intruder will react to certain defense systems and network architectures, allowing administrators to reduce the damage caused by individual attacks. The data collected through the process can also assist in predicting future cyber intrusions.
“The implementation of intrusion prediction into existing cybersecurity systems is believed by many experts to be the next step in elevating network defense,” notes Shanchieh Jay Yang, assistant professor of computer engineering at RIT. “By modeling how a hacker will act within a network, administrators can better create protocols to defend essential data and reduce the impact of the intrusion to the overall network.”
The project is being funded under a sub-contract from the CUBRC/University of Buffalo Center for Multisource Information Fusion. The team also includes personnel from the U.S. Air Force Research Laboratory’s Information Fusion Directorate, which hopes to use the research to assist in implementing intrusion prediction into their overall cybersecurity systems.
“The addition of these advanced predictive models will enable the system being developed on this contract to anticipate attacks for network analysts,” states Moises Sudit, professor of industrial and systems engineering at the University of Buffalo and managing director of the Center for Multisource Information Fusion. “It will also provide critical input to other models contained in the system that are designed to automatically protect the network when an attack occurs.”
The research team recently tested an initial prototype of the system and plans to begin further implementation over the next year. They are also looking to develop similar intrusion prediction models for use in additional military and commercial networks.
“Given the growth of the Internet and computer networks over the last 15 years and their importance in nearly every facet of American life, the security of the system is a major issue for scientists, government, industry and consumers,” says Yang. “It is my hope the use of intrusion prediction will improve network defense and ultimately lead to better forecasting and prevention of future attacks.”