After almost two decades of research on anonymous network communication the development has forked into two main directions, namely Mix cascades and peer-to-peer (P2P) networks. As these design options have implications on the achievable anonymity and performance, this paper aims to elaborate the advantages and disadvantages of either concept. After clarifying the scope of the discussion, we present arguments for Mix cascades and P2P designs on multiple areas of interest: the level of anonymity, the incentives to cooperate, aspects of availability, and performance issues. Pointed thesis and antithesis are given for both sides, before a final synthesis tries to articulate the status quo of the discussion.