This book is the result of 18 years of research into Steiner's problem and its relatives in theory and application. Starting with investigations of shortest networks for VLSI layout and, on the other hand, for certain facility location problems, the author has found many common properties for Steiner's problem in various spaces. The purpose of the book is to sum up and generalize many of these results for arbitrary finite-dimensional Banach spaces. It shows that we can create a homogeneous and general theory when we consider two dimensions of such spaces, and that we can find many facts which are helpful in attacking Steiner's problem in the higher-dimensional cases. The author examines the underlying mathematical properties of this network design problem and demonstrates how it can be attacked by various methods of geometry, graph theory, calculus, optimization and theoretical computer science. Audience: All mathematicians and users of applied graph theory.