Complex Event Processing (CEP) Glossary
To better define Complex Event Processing for real-time stream processing, here is a concise glossary:
Tuple
A tuple is a collection of name-value pairs. The permissible names and the types of the values are expressed in a schema definition which specifies the "type" of a tuple. Tuples may be "flat" if the values are drawn from the set of atomic types (e.g., integer, float, string) or they may be "hierarchical" if the values can be drawn from structured types (e.g., set, tuple, list).
Stream
An ordered and potentially infinite collection of tuples. The ordering is on arrival order at a given processing input, and this ordering may or may not correspond to other related orderings (e.g. tuple creation order, wallclock, time of a tick). Such additional orderings are not required. Also, it should be noted that operating on one or more streams that are ordered by some attribute may produce an output that no longer retains that order.
Stream disorder
The tuples in a stream may or may not be in order by some desired attribute of those tuples. When such an ordering exists, some computations become easier and can be performed without the need for arbitrary storage; however, when this ordering is violated, this is called "stream disorder."
Stream processing
Stream processing is a technology that facilitates low latency filtering, correlations, aggregations and computations on streams, so as to enable real-time actions or alerts. A stream processing system must be prepared to produce results even when stream disorder exists. Stream processing systems also must combine stream data with stored data in order to efficiently produce results that can depend on both real-time events and historical data.
Event
An event is an interpretation of a tuple. While a tuple is simply a data structure, an event corresponds to something in the real world. For example, if a blue car is observed at location (100, 100), then the tuple (object-type: "car", color: "blue", x: 100, y:100) might be a good representation for that event.
Event processing (also known as "complex event processing")
Event processing refers to filtering, correlating, aggregating, and otherwise computing with tuples that represent real world events. These computations typically produce other events that can be further processed.
Poset processing
Poset processing consists of performing computations on a set of tuples that may not be related by a total ordering. Any partially ordered set of tuples can be processed in arbitrary ways within a stream processing system by storing those tuples and retrieving as needed to match desired patterns. These patterns can be expressed in any language (i.e., regular expressions, SQL).
« Previous 1 | 2