UC Santa Barbara

Current large-scale IPv6 studies mostly rely on non-public datasets, as most public datasets are domain specific. For instance, traceroute-based datasets are biased toward network equipment. In this paper, we present a new methodology to collect IPv6 address datasets that does not require access to restricted network vantage points. We collect a new dataset spanning more than 5.8 million IPv6 addresses by exploiting DNS' denial of existence semantics (NXDOMAIN). This paper documents our efforts in obtaining new datasets of allocated IPv6 addresses, so others can avoid the obstacles we encountered.