RGen: A Real-Time Pattern Mining Approach to Target Generation for Internet-Wide IPv6 Scanning

Target generation is a crucial step for efficient Internet-wide IPv6 scanning, yet existing techniques are hindered by limited pattern discovery and biased target distribution. We introduce RGen, a novel target generation algorithm employing real-time pattern mining and a stochastic generation strategy. RGen minimizes pattern loss, dynamically incorporates new addresses, and promotes balanced target distribution, overcoming prior limitations. Experiments show RGen achieves the highest overall hit rate (58%), surpassing the previous best (AddrMiner-S) by 60%, while simultaneously leading in network discovery, finding the most new BGP prefixes and new / 64 prefixes.