Active database systems extend traditional database systems with the capability to react to various events, which can be either generated internally or externally. We discuss the event‐condition‐action (ECA) paradigm to specify reactive behavior and its enabling tools and triggers which are available in commercial DBMS. We motivate the need for formal reasoning about active database behavior and present the desirable global properties of termination and confluence. Then we discuss the various semantic dimensions of triggers that relate to their instantiation, condition evaluation, scheduling, and execution and we discuss their coupling with the underlying transaction model. The main features of prototype active database systems and of commercially available DBMS are reviewed. We introduce some of the challenges that face active database systems in novel, highly heterogeneous, distributed, and dynamic driven application domains such as workflow systems, streaming database, and moving objects databases. Finally, we give a brief overview of a new paradigm, (ECA)'2, which is geared toward settings in which both the event detection and the condition evaluation are done in distributed fashion and are continuous in nature.