As part of Risk Management, there are event detection techniques that make it possible to anticipate the occurrence of incidents.
It should start with the selection of business and operational risk detection metrics.. From these defined metrics it is feasible determine alarm mechanisms.
ETL tools can examine data sources in batch mode or near real time looking for criteria in the data. These criteria can be:
- simple, how to exceed a certain value,
- complex, that require the construction of a decision mesh.
Faced with this second case, a mixed model is proposed:
- Data update frequency varies in each system.
- The decision mesh can have a first decision cycle. based on low refresh rate indicators that show partial results.
- The mesh is completed with the indicators obtained through rapid update processes. and the end result is an alert semaphore.
Decision meshes can examine databases, flat files and contain complex logic as a search for specific values and be the result of table crossovers, counts and consultation of historical information.
Support in a system with integrated and standardized data is of great help in this procedure, although It may be necessary to make direct inquiries to Transactional Companies for the detection of events of rapid update and high criticality..
Results can be reported passively or electronically instantly (email, messages, control boxes, etc.).
