QONCEPT

Semantic Query Optimisation in Complex Event Processing Technologies by Constraints on Event Streams

The project QONCEPT develops and evaluates methods for query optimisation in Complex Event Processing.

Complex Event Processing, short CEP, denotes algorithmic methods for making sense of events by deriving higher-level knowledge, or complex events, from lower-level events in a timely fashion and permanently. Examples are the recognition of trends in markets, of situations that might become difficult or dangerous in transport systems, or of risks of failures in industrial or manufacturing processes.

At the core of CEP are queries incessantly monitoring the incoming stream of "simple" events and recognizing "complex" events from them. Event queries serve the specification of situations that manifest themselves as certain combinations of simple events occurring, or not occurring, over time and that cannot be detected from each single event involved.

Querying events is fundamentally different from traditional querying in databases or querying of Web data, since event queries are standing queries evaluated permanently over time against changing event data received as an incoming stream. Special purpose Event Query Languages (EQLs) have been developed for the expression of the complex events in a convenient, concise, cost-effective and maintainable manner. Specific evaluation methods have been conceived for the efficient, stepwise evaluation of complex event queries against event data streams.

In the project QONCEPT, properties of an incoming event stream will be used for optimising the evaluation of complex event queries against this event stream. Properties of several kinds will be considered:

• restrictions of the kind of simple events received,
• structural properties of the event stream, i.e. dependencies such as a delivery always follows an order, and
• temporal properties

Such properties are present in most of the complex event processing applications. Similar properties have proven to be very useful in optimising the evaluation of standard queries. Their use for query optimisation in CEP is therefore promising.