Analysis And Measurement Of Zone Dependency In The Domain Name System

The Domain Name System (DNS) is a hierarchical distributed system organized through top-down zone delegation. Consequently resolution of a zone depends on its ancestors. However, since the delegation in DNS is designed by name rather than address, the dependency could further extend to other zones. If not configured well, the dependency of a zone could be large and complicated, potentially harmful to its availability and integrity. In this paper, we propose a graph-based model to comprehensively analyze zone dependency in DNS. Our approach classifies zone dependency into four different relations: general dependency, explicit dependency, critical dependency and essential dependency. We also propose an empirical method to quantitatively measure the zone dependencies of given zones. Our survey with over 1 million DNS zones shows that more than 99% of the zones depend on some 3-rd party zone; about 41% of the zones critically rely on more than 2 zones except their ancestors; some TLDs such as.org,.info and .cn tend to have more dependencies than others.