Stream replication uses the information available in redo logs to synchronize data between two servers. The capture process captures the changes at source, and the changes are transported to the target database, which generates the transactions and the apply process applies those transactions.