Attack Trace Dataset Generation Considering Organizational System Characteristics

Masahito Kumazaki, Daisuke Kotani, Yasuo Okabe

Abstract: In the context of cyberattack response, leveraging damage states that can potentially be reported earlier than data requiring specialized analysis may lead to reduced response times and mitigated damage. However, while datasets such as those for attack sequences exist, there is a lack of datasets that correlate attack methods with their respective damage states. This paper proposes a dataset generation tool for attack traces, including mappings between attack methods and damage states. The proposed tool operates within a virtual environment that replicates organizational settings, automatically executing attacks and collecting information. This approach addresses challenges such as variations in data caused by different environments and the burden of manual data collection.

External IDs:dblp:conf/icoin/KumazakiKO25