Monitoring applications periodically acquire data, store them in the database, and perform data analysis to detect the events in the environment. The applications repeatedly poll the database to check that new data has arrived. High poll rates waste significant resources, while low poll rates result in slow response and risk missing a sudden change. Both approaches lead to unsatisfactory performance of applications.

What is the cause of the problem? Conventional databases act as passive components incapable of initiating processes on their own. Although the DBMS are cognizant of their state's change, many of them lack the functionality necessary to make the client aware of this alteration. The databases with this essential functionality are called active databases (ADBMS).

For monitoring applications, ADBMS are a natural fit. They improve application performance by guaranteeing almost instantaneous reaction to database state change without wasting many resources. However, their biggest impact is in enabling a new type of system architecture that brings new capabilities to data analysis and decision making processes.

We start the presentation with discussion of several types of monitoring systems. For each system we will compare conventional and new architecture and show how an active database improves client performance and functionality. Then we will demonstrate that ADBMS adds completely new capabilities to the database server as well. For example, they allow an automatic search for predefined event patterns in the acquired environmental data on the server. The search is executed by the Rules Engine component, which is designed to discover the relations between seemingly unconnected events.

The ability to perform automatic event pattern identification elevates monitoring systems from a mere registration device into a powerful controlling tool. By being able to find relations between events which do not have obvious connections they are capable of recognizing underlying activities and preventing undesirable effects by influencing the decision process.

Once the new generic monitoring architecture is analyzed, we will discuss the theoretical foundation upon which the Rules Engine is based. We will explain the theory of events, event-condition-action rules and how these rules are used to identify event patterns.

In the last part of presentation we will show that the latest version of MySQL has all necessary features to qualify for an active database title.