Traceability Measurement Analysis of Sustained Internet-Wide Scanners via Darknet

Darknet observation networks have experienced a significant surge in sophisticated and diverse Internet-wide scanning activities and investigative scanner operations. In addition to malware scanning attacks, reconnaissance scans to routinely probe the Internet space are also increasing in size to the point where they interfere with darknet data analysis. Since scans for investigative purposes that are not directly related to threats cause false positives in analysis, we desire to segregate them and focus on more substantive attack activities. Recent studies have proposed methods for analyzing these investigative scanners by clustering those with similar characteristics and tracing clusters based on IP addresses. However, these studies have primarily concentrated on specific cases and have yet to examine the traceability of all Internet-wide scanners comprehensively. In this study, we implement a naive cluster tracing method similar to those in existing studies to evaluate traceability. Specifically, we cluster scanners daily based on three features: the number of packets, the number of destination ports, and the number of destination IP addresses, and we assess traceability using the change rate of IP addresses and the distance of centroids. Our results reveal that approximately 98.4% of the scanners are difficult to trace, highlighting the need for more advanced and complex tracing methods.