U.S. Army researchers have developed a new machine learning-based framework that optimizes the moving target defense method used in protecting in-vehicle computer networks.
The deep reinforcement learning-based resource allocation and moving target defense deployment framework helps identify the optimal internet protocol shuffling frequency and bandwidth allocation for effective moving target defense, the Army said Tuesday.
Frederica Free-Nelson, an Army computer scientist and the DESOLATOR program lead, noted that the framework keeps networks unpredictable for hackers without being costly for operators and affecting computer performance.
By fastly shuffling IP addresses, "the information assigned to the IP quickly becomes lost, and the adversary has to look for it again," explained Terrence Moore, an Army mathematician.
The Army Research Laboratory (ARL) teamed up with Virginia Tech, Australia's University of Queensland and South Korea's Gwangju Institute of Science and Technology to develop DESOLATOR.
Since it is based on machine learning, the framework can be modified by other researchers to solve various problems.