Research on reconfigurable intelligent surfaces (RISs) has predominantly focused on purely physical (PHY)-layer aspects, particularly, on how signals are dynamically shaped by a controllable wireless propagation environment. However, integrating RISs as system-level network elements requires the development of an RIS-compatible control plane. In this article, we explore design options for such a control plane across two key dimensions: the allocation of spectral resources for the control plane (in-or out-of-band), and the rate selection for the data plane (multiplexing or diversity). While our analysis is necessarily simplified, it reveals the fundamental trade-offs inherent in these design choices, which are crucial for integrating RIS technology into future networks.