• Game Experience Management - A client management application. It would use transient game events to build a picture of the current client experience and  tracking promotion drive impact.
• Financial Fraud Detection - Event processing over sliding window of time can used to detect subtle pattern in client behavior.
• Process Supervision - In large processes correlation of individual modules may present a non-trivial problem. Using CEP problem areas can be detected earlier and a large scale picture of the system health can be updated in real-time.

Complex Event Processing

Complex Event Processing, or CEP, is primarily an event processing concept that deals with the task of processing multiple events from a so called event cloud with the goal of identifying the meaningful events within the event cloud. CEP employs techniques such as detection of complex patterns of many events, event correlation and abstraction, event hierarchies, and relationships between events such as causality, membership, and timing, and event-driven processes.

Significant and Non-Significant Events

CEP differs from traditional data warehousing in that it is based on system events which may not be feasible  to write to database. For example, it is not uncommon for large game systems to process several thousand events per second, but only a subset of which are directly significant. In such case CEP is used to aggregate non-significant event into significant ones.

•  A non-significant event is a system event that is not interesting immediately, unless correlated with other events and system data. For example, "client A withdraws money" is not significant, however "client A withdraws money 3 times within a minute, and his account is simultaneously filled up by user B, who is a new user in the system", which is a combination of multiple events within a time frame and latent system knowledge, might be.
• When several non-significant events are combined with facts and optionally a sliding time window, one or more significant events may be produced which describes the triggering process. Given the example above, a "money laundry warning flag" might be raised and passed on in the system. 

Staged Processing

A complete CEP installation is usually modeled like this:

• Events are produced from system. This can be transient events within an "event cloud" in a distributed system, events reaped from log files or triggers acting directly on user input.
• The events are aggregated and processed in on or more CEP stages from which significant events are produced and passed on. The significant events may also be passed through a series of further CEP stages for additional processing.
• The significant events are optionally passed to a customer UI for direct presentation and monitoring.
• Finally the significant events are stored in a data warehouse.

Below is an example flow chart of game experience manager installation:

Articles

• How to Implement a Cubeia CEP Installation
• Real Time Business Intelligence

Implementations

• Cubeia GEM - Complex event processing for the gambling industry.

More Information

For more information, do not hesitate to contact us.