In distributed model predictive control (DMPC) complex large scale systems are decomposed into a set of smaller subsystems. These subsystems utilize local model predictive controllers and rely on communication between the controllers to achieve good control performance. This often results in a very high load on the communication network. Thus, a key issue is how and when to communicate information between the controllers. The topic of this talk are recent results on event-based communication in DMPC, which allow balancing the system performance and the load on the communication network.