Discovery of IPv6 router interface addresses via heuristic methods

Abstract

With the assignment of the last available blocks of public IPv4 addresses from Internet Assigned Numbers Authority, there is continued pressure for widespread IPv6 adoption. Because the IPv6 address space is orders of magnitude larger than the IPv4 address space, researchers need new methods and techniques to accurately measure and characterize growth in IPv6. This thesis focuses on IPv6 router infrastructure and examines the possibility of using heuristic methods in order to discover IPv6 router interfaces. We consider two heuristic techniques in an attempt to improve upon current state-of-the-art IPv6 router infrastructure discovery methods. The first heuristic examines the ability to generate candidate IPv6 addresses by finding the most common lower 64 bit patterns among IPv6 router interface address observed in historical probing data. The second heuristic generates candidate IPv6 addresses by assuming that an IPv6 address seen in historical probing data is one end of a point-to-point link, and uses the corresponding end’s IPv6 address. Using a distributed active topology measurement system, we test these heuristic methods on the IPv6 Internet. We find that our first heuristic is successful in discovering a non-trivial number of new router interfaces, while the second heuristic is more efficient.