Scan engines are the Scuba processes that analyze streams of data during scans. Scan engines come in two flavors: aggregation and annotation.
Aggregation engines apply a function that results in a single value.
Annotating engines result in an array of values where every array row is associated with a row of data in the analyzed time interval.
These annotations can be treated as virtual columns that enrich the recorded event data and either contain the desired query results or are used as input for further passes over the data.