This paper addresses the challenge of constructing center specific multicast trees in a global network. The described approach is an interaction specific algorithm for data distribution center location using asymmetric traffic loads on network links. A fast and scalable algorithm for locating distribution centers based on the network load and participant's locations is given. This algorithm is presented as part of a tool which interactively allows a user to build networks, specify the group of multicast participants, select manually or generate algorithmically a center for the multicast tree and evaluate the tree quality. Simulation results on various topologies are presented showing that, with the above center location mechanism, algorithmically located center-specific trees yield lower tree cost and delay than center-specific trees with randomly selected centers.