TracerForests6: Efficient IPv6 Alias Resolution via Traceroute Path Expansion

Abstract

In recent years, IPv6 networks have seen rapid global deployment. Router-level topology is essential for understanding network structure and supports performance monitoring, security analysis, and fault diagnosis. A key step is alias resolution, which maps multiple IP addresses to the same router so that observed interface addresses are correctly attributed. However, the vast IPv6 address space presents new challenges for topology discovery. Traditional IPv4 approaches rely on exhaustive collection of router interface addresses followed by alias resolution. Many IPv6 methods inherit this approach and suffer poor scalability, limited efficiency, and reduced coverage at IPv6 scale. In this work, we present TracerForests6, a heuristic and efficient method for IPv6 alias resolution. It continuously discovers alias pairs across the network starting from high-confidence initial seed pairs. It analyzes traceroute paths to the initial set to identify candidate pairs from path characteristics, and a random forest trained on multidimensional path features validates them. In our evaluation on CAIDA ITDK and two RIPE Atlas traceroute snapshots, using the initial seed construction and 14-vantage-point probing configuration adopted in this work, TracerForests6 expands the identified alias set to $1.78\times $ the corresponding initial seed set while maintaining an alias-pair identification accuracy of 96.85%. To facilitate reproducibility and follow-up research, we will release a set of public resources related to TracerForests6, including the implementation and measurement artifacts, which can support further IPv6 topology analysis.