How High-speed Apply Engine works

The following figure shows an overview of the High-speed Apply Engine data flow. The numbered steps for the data flow are described following the figure.

When you submit an apply request, High-speed Apply Engine processes it as follows:

1. The request processor manages the apply request.

2. The request processor also obtains configuration information for the apply request and determines the processing requirements from that information. You can specify configuration information in the apply request or in a separate configuration file.

3. The input processor begins processing the input source that is specified in the configuration information.

4. If the apply request specifies restart processing, the input processor uses information in the restart table and the configuration file to determine the location of the first record to be processed in the input source.

5. The input processor passes SQL statements to an apply agent. Each agent processes the statements for an object, set of objects (tables or partitions), or a unit of recovery (UR). You specify the maximum number of agents in the configuration information.

6. The apply agents execute the SQL statements against the target tables.

7. The apply agents keep track of the changes that have been applied, display statistics about the request, and update the restart table.

8. If conflicts occur during apply processing, the apply agent coordinates conflict resolution with the conflict manager. The conflict manager uses the conflict resolution rules in the configuration to determine the appropriate action.

9. If you configure the apply request to include defer actions and a conflict file, the conflict manager writes the deferred conflicts to the conflict file.