Mass Ingestion Databases provides a single change capture mechanism and architecture for capturing change data from a Db2 source on IBM i and applying the data to a target.
The Secure Agent must run on a Linux or Windows box, separate from the IBM i source system. The network bandwidth between the IBM i system and the Secure Agent box must be robust. Informatica recommends network transfer rates that can handle100s of gigabits or one or more gigabytes of log data. If the network transfer rate is not capable of delivering the log data to the Secure Agent at speeds equal to or greater than the rate at which Db2 produces log data of CDC interest, the database ingestion job will not be able to provide the data to the target in a timely manner. If the data throughput does not meet your SLA, consider changing the hardware to increase the ethernet bandwidth between the IBM i system and the Secure Agent box.
The following image shows the Db2 for i change capture components and data flow:
The log collector, which runs in the DBMI Agent service under the Secure Agent, sends a request to the Db2 DISPLAY LOG command for Db2 for i journaled data.
Each request includes a starting RBA along with a list of tables of CDC interest for database ingestion.
The command requests the data for the source tables from the Db2 journals and journal receiver.
The command returns the journal entries containing the data to the UOW (Unit Of Work) Manager.
The UOW Manager sends the journaled data in committed transaction order to the log parser.
The log parser parses the DML changes from the committed transactions. Before sending the data to the CDC writer, the log parser transforms it into a canonical form of the Db2 journal data that can be consumed by the DBMI Agent task and applied to the target.
Because this resource-intensive activity occurs on the Secure Agent box, CPU consumption on the IBM i system is minimized.
The CDC writer applies the formatted data to the target.
ASK INFAPreview
