Glossary

Bobo Events

Data types used in bobocep to represent different system events.

PrimitiveEvent

Represents data that has entered the bobocep system.

CompositeEvent

Represents the inference of a complex event.

ActionEvent

Represents the execution of an action by the system.

Complex Event

An event that represents the inference of some phenomenon that was identified by a pattern in a data stream.

Contiguity

The policy of states with regard to how they react to events that do not cause a state transition. bobocep supports the following three types of contiguity.

Strict

All matching events are strictly one after the other, without any non-matching events in-between. If an event does not match, the run halts.

Relaxed

Any non-matching events are ignored.

Non-Deterministic Relaxed

The same as relaxed, but allows multiple matches from a state when its transition is non-deterministic.

History

The events that were accepted by a pattern as being indicative of the existence of a complex event.

Null Data

Arbitrary static data that is periodically inserted into the CEP system. It has several purposes:

• Provides state clearance, so that runs always have a periodic event with which to trigger time window checks.

• So runs can reach the accepting state if nothing happens i.e. check if a null event occurs instead.

Pattern

A sequence of data correlations that, when fulfilled with data from a data stream, infer the existence of a complex event.

Run

An instance of a pattern.

State Clearance

Ensuring that runs are always eventually halted and removed from memory, to prevent a build-up of incomplete runs with no means of halting. This is typically achieved by adding a time window to patterns that will cause an eventual halt if a run is not fulfilled within a given time frame.

States

Start State

The first state of an automaton that triggers the generation of a new run when reached.

Accepting State

The final state of an automaton that triggers the generation of a complex event when reached.