I. Introduction
The rising relevance of Electromagnetic Compatibility (EMC) issues in technical applications increases the importance of numerical simulations. The physical laws of high-frequency (HF) electromagnetic problems can be described using Maxwell's curl equations. Several discretization methods like the Finite Difference Method (FD), Finite Volume Method (FVM), Finite Integration Technique (FIT) or the FEM exist to allow for numerical simulations of Maxwell's equations. Along these techniques, DG methods, which are a variation of the conventional FEM, have gained attention, i.e. in [1], [2], [3]. In the EMC-society, the DG-Method was recently introduced as an efficient, high-order and highly parallel alternative to other techniques [4]. The spatial discretization is based on unstructured, polyhedral finite elements with local ansatz-and testfunctions, being independent to those of connecting elements. As a consequence, numerical solutions are allowed to be discontinuous along element boundaries. Neighboring elements couple through numerical fluxes. These fluxes are responsible for connecting the local solutions on the elements to one global solution on the computational domain. Furthermore, the fluxes control the conversation of physical boundary conditions, i.e. tangential continuity of the electric and the magnetic field, respectively.