Capturing change data with single scrape

Single scrape, also known as shared scrape, is a source-only component of CDC Replication that allows multiple subscriptions in an instance to share the same log reader and log parser thread.

With single scrape, CDC Replication only reads and parses the source database transaction log once to capture changes for all tables being mirrored for the instance.

Single scrape improves performance by reducing the footprint on your source system since it only requires a single log reader thread and a single log parser thread to service multiple subscriptions. This reduces disk I/O and decreases CPU utilization by CDC Replication processes.

Change data and the single scrape staging store

After the CDC Replication log reader captures the change data from the database logs and the data is parsed by the CDC Replication log parser, change data is placed in the staging store. Each subscription retrieves the changes for mirroring tables from the staging store whenever possible. The data in scope for a given subscription is kept in the staging store until it is sent to the target database. After the data is sent to the target it is removed from the staging store. To improve performance when subscriptions are mirroring, CDC Replication will keep the staging store data in memory whenever possible.