Charles River Analytics Develops MAAGI to Support DARPA's Cyber Genome Program
The Defense Advanced Research Projects Agency (DARPA) created the Cyber Genome program to combat the growing threat of cyber attacks on US resources. As part of the Cyber Genome program, Charles River Analytics Inc., a developer of intelligent systems solutions, developed and is refining MAAGI (Malware Analysis and Attribution using Genetic Information). In its current version, MAAGI combines ideas and techniques from biological evolution, reverse engineering of computer programs, and linguistics to rapidly identify the source and intent of new malware attacks. Our MAAGI case study appears below.
Cyber attacks, such as viruses, Trojans, and worms, are a growing threat to US missions and resources. To combat the growing threat of cyber attacks on US resources, DARPA created the Cyber Genome program. Cyber Genome aims to develop revolutionary, new cyber-forensic techniques to automate the discovery, identification, and characterization of malware variants.
The Charles River Analytics Solution
As part of the Cyber Genome program, Charles River developed and is refining MAAGI. In its current version, MAAGI combines ideas and techniques from biological evolution, reverse engineering of computer programs, and linguistics to rapidly identify the source and intent of new malware attacks. MAAGI makes use of the fact that malware authors often reuse code from one attack to the next, while trying to conceal this reuse from defenders by changing the “surface” features of the malware. By discovering the essential “genetic” properties of malware that are preserved from one malware sample to the next, MAAGI seeks to determine the lineage of each sample and uses the lineage to help characterize the source of the malware.
Furthermore, by understanding the patterns of evolution in malware, MAAGI can be used to predict future malware development, anticipating potential attacks rather than — as we do today — merely reacting to them. MAAGI also uses methods from functional linguistics to identify the functional features and potential intent of malware, aspects that are especially likely to be preserved even when surface features change. MAAGI allows an analyst to view the evolution of malware on a gene-by-gene basis.
MAAGI is an innovative approach to the Cyber Genome challenge of characterizing and predicting the evolution of malware. It supports detection and attribution of cyber attacks for both the defense and law enforcement communities.
By recognizing code and techniques from previous attacks, MAAGI enables quicker response times to defend against cyber attacks. MAAGI is proactive in that it not only assesses attacks, but anticipates and predicts the properties of future attacks. Finally, MAAGI changes the economics of malware by making it more difficult for malware authors to change superficial features and reuse their code.
Source : Charles River Analytics