In the IPv4 world the IP parameters of servers (like their addresses, DNS resolvers and NTP servers to use) were usually assigned in a static manner, either manually or by means of automation. Taking a look at some main architecture elements of IPv6 it becomes clear that this is not always the most efficient approach with IPv6. In this talk I will discuss potential approaches and their respective advantages/disadvantages, based on deployment experience in very large data centers. Also I will cover which individual properties of some common operating systems lead to specific constraints or configuration requirements as for the network infrastructure in place.