An IPv6 Target Generation Approach Based on Address Space Forest

IPv6 target generation techniques are crucial for Internet-wide rapid scanning of IPv6 network assets. Current algorithms are mostly limited to low-dimensional patterns (pattern dimensions ≤ 4) within the IPv6 address space tree (6ASTree). Due to the uneven distribution of IPv6 seed addresses and irreversibility of clustering in existing IPv6 target generation algorithms, the number of low-dimensional patterns in single-tree algorithms like 6Scan and 6Tree, or dual-tree algorithms like HMap6, is limited. Additionally, the large-scale loss and high-dimensional pattern spaces pose challenges of undetectability and excessively large probing space. To address these issues, an IPv6 target generation approach (6Probe) based on IPv6 address space forest (6ASForest) is proposed. By constructing multiple 6ASTrees to form a 6ASForest, 6Probe leverages seed address structure to explore high-activity regions in both high-dimensional pattern and loss spaces, significantly expanding the scale of IPv6 active addresses probed. On balanced IPv6 seed set C5, 6Probe can probe 2.6 to 4.81 times active addresses, and produce 5.47 to 20.01 times low-dimensional nodes compared to existing 5 typical algorithms (6Scan, HMap6, 6Tree, 6Gen, 6Hit). On unbalanced IPv6 seed sets, 6Probe can probe 137.87% to 492.74% of active addresses, and produce 430.81% to 712.80% of low-dimensional nodes compared to the HMap6.