Correlation engines are the underlying processes that drive ALEC. Two engines are implemented by default: clustering and deep learning. Both use the network topology graph that is built and maintained during pre-processing, based on information provided by your datasource (see ALEC Datasources).
Before information is passed to a correlation engine, ALEC builds a network topology graph. The graph is based on network inventory information that the datasource provides, and is persisted in local memory. Inventory objects are added to the graph as vertices, and relationships among them are added as edges. Alarms are then attached to the vertex of the inventory object that they reference.
On every refresh interval (which defaults to 30 seconds), ALEC runs a clustering algorithm against all alarms on the graph. This process outputs a list of alarm clusters. The engine then maps clusters to situations, and sends that situation data back to the datasource.
Pre-processing functions handle all aspects of interfacing with the datasource and managing state information.