CyberARAIM: A Model for Rapid and Precise Identification of Affected Digital Assets in Practical Combat Scenarios with Analysis of Boundary Scope and Dependency Relationships

In recent years, cyberspace attack and defense confrontations have become practical and routine. On one hand, cyber attackers are employing increasingly covert and diversified methods and tools with more frequent attacks, while on the other hand, defenders need to protect an ever-growing and ubiquitous array of digital business operations and assets, referred to as "digital assets." These assets are characterized by extensive ubiquity, exposure, and dependency. Facing systematic confrontations and systemic risks that persist over time, the ability to rapidly and precisely identify the boundaries and interdependencies of affected digital assets among tens of thousands, hundreds of thousands, or even millions of dynamically operating digital assets is crucial. This capability supports the prompt and accurate positioning and orderly, scientific handling that are fundamental to security defense efforts.The CyberARAIM model, refined from years of practical experience, scenario verification, and iterative applications, utilizes multi-source dynamic collection of basic digital asset data and trusted data governance to create a knowledge graph of digital assets. Based on this graph, the model applies the BERT algorithm combined with a decision tree algorithm informed by expert experience for analysis and identification. In practical engineering applications, this approach has significantly improved the speed and precision of locating the boundaries and dependencies of affected digital assets, reducing the time required from several days to one minute, with an accuracy rate exceeding 94%. This paper completes the model research, experimental verification, and engineering application.