Neta Peleg

We investigate the negative caching (caching of NXdomain
responses) behavior on nine large open DNS resolvers. We
measure the amount of time an NXDomain response is kept
in the cache in various TTL configurations and compare it
to the time an existent domain is kept in the cache.

In this work we measure what percentage of DNS recursive
resolvers perform negative caching in the wild. We deploy
our own authoritative name server and harness thousands
of RIPE Atlas [3] sites spread over the globe to perform
repeated DNS queries for non-existing sub-domains of our
authoritative domain.