This paper proposes to use optimization and machine learning methods in order to develop innovative techniques for balancing the DNS(SEC) traffic according to Fully Qualified Domain Names (FQDN), rather than according to the IP addresses. With DNS traffic doubling every year and the deployment of its secure extension DNSSEC, DNS resolving platforms require more and more resources. A way to cope with these increasing resources demands is to balance the DNS traffic among the DNS platform servers based on the queried FQDNs. Several methods are considered to build a FQDN based routing table: mixed integer linear programming (MILP), a Kmeans clustering algorithm and a heuristic scheme. These load balancing approaches are run and evaluated on real DNS traffic data extracted from the operational IP network of an Internet Service Provider (ISP) and they result in a difference of less than 2% CPU between the servers of a platform.