Time is the shadow of reactive behaviour

Applications that rely on coordination of messages are frequently based on multi-agent systems or workflow systems. Both implementation platforms use a rule engine for the coordination of messages. Currently, classical production rules are used within multi-agent systems, and workflow systems tend to rely on active database solutions (i.e. triggers). It has been envisioned that subsequent application generations are likely to require support for hundreds or even thousands of triggers. This is in contrast to the current state-of-the-art of implementations, which only scale to a few triggers. The paper outlines a vision where a scalable trigger system is the unifying concept between workflow and multi-agent approaches for applications that require coordination facilities. In particular, the paper defines scalability and performance within an active database context. The paper explores factors and situations that influence active database performance. Finally, the paper explores promising directions for how to move the state-of-the-art to trigger systems that scale to many triggers.