A patent involving Florida Tech faculty may protect growing computer networks with ever-growing cybersecurity threats.
Marco Carvalho, Florida Tech executive vice president and provost, and computer engineering and sciences associate professor Thomas Eskridge, working in collaboration with engineers from Raytheon BBN Technology Corp., have developed a “decision engine” to configure the best set of defenses to combat threats against a system. The engine uses a genetic algorithm to search through several possible defense configurations and has an interface where users can interact and provide feedback on the direction and characteristics of the evolved solutions. The patent was awarded December 2020. The team also developed a patent-pending user interface that allows the operator to direct the genetic algorithm.
By using a genetic algorithm, the system can decide where to place defenses on the network based on the user’s preferences, allowing for stronger and more efficient protection than seen in previous systems. Eskridge noted the team tested the application by using over 100 possible placements and configurations of defenses, with a set of criteria that evaluates how well they perform. The results led to successful deployments via the algorithm’s adaptability in providing defense against attacks.
“I know what a good defense system will do for me: It will block these bad things, it will let my traffic through, it will keep privacy on this particular link at a very high level. It will do all these things. I just don’t know how to configure it to do it,” Eskridge said. “So, the algorithm figures out if a particular configuration is what you’re looking for and provides the best way to get what you want for as little cost as possible.”
The work on this technology is related to the federated security research Florida Tech has been involved with since 2013, led by Carvalho as principle investigator.
What started out as research on what basic infrastructure requirements would be needed for cybersecurity evolved into research focused on a combination of large network communications, decision algorithms and data visualization tools. Florida Tech’s research recently led to security demonstrations that featured a network of 1,100 computers.
“We are building intelligent algorithms capable of optimizing complex cyber defense infrastructures on demand, and in collaboration with human analysts,” Carvalho said.
The security customization sparked by the patented algorithm also allows for much more user flexibility by researchers. If the algorithm finds a solution that can better protect systems from threats, Eskridge, Carvalho and their team are then able to make variants of the code for different parts of the system.
While this research is mostly for government use, there is a version the team is working on that allows people to share information with each other and to automatically enhance their security. For example, it tracks suspicious IP addresses, avoids them and shares them with users on the network, giving all users an extra level of security. A demonstration of this system will take place in the coming weeks.
The ability to continue to optimize network defenses is key in a world that, according to Security Magazine, sees 2,200 cyberattacks per day, or approximately one every 39 seconds.
“I think this will really be useful for home and small business, just because it’s difficult to keep up with all the updates,” Eskridge said. “If our system can do this for your automatically, we’re going to make people much, much safer.”
Both Carvalho and Eskridge are members of the National Academy of Inventors (NAI), a member organization comprising U.S. and international universities, and governmental and non-profit research institutes, with over 4,000 individual inventor members and Fellows spanning more than 250 institutions worldwide. The NAI publishes the multidisciplinary journal Technology and Innovation, Journal of the National Academy of Inventors.