The Producer and Consumer Model

For instance, in a source database the table ‘orders’ is updated by application transactions. This action generates redo entries into the redo logs. The changes are produced at the source database and this database becomes the producer. These transactional changes are then routed to a destination database where the Streams client, via the Apply process, consumes them to update the orders table.  The destination database becomes the consumer. There could be even multiple consumers especially when the changes are routed to multiple destinations.

Thus, the consumer is typically the destination database in which database changes or user messages that have been system generated by the producer are ultimately consumed. Once changes are generated at a source database, they may be consumed by more than one remote database.  In addition, any application can consume messages. This model of “producer and consumer” controls the nature of the data flow origin and ultimate usage.

Streams Clients and Event Flow

In the Streams environment, there are four distinct entities which act as the clients. They are called Streams Clients. These distinct entities are: the Capture process, the Apply process, the Propagation process and the Messaging client which is typically a user application. All operations are conducted by these four entities, and they act as the producers and consumers. Any activity for data flow within Streams is controlled and managed by these four Streams Client processes. They are typically the Oracle processes which perform the data flow.