[Oracle] Redo transmission of Data Guard

Data Guard provides two services:
1) Redo transmission service: Transfers Redo logs of Primay to one or more Standby destinations.
2) Redo Application Service: The Redo log transmitted from Primay on the Standby side.
This article describes the Redo transmission service.

1. Use ARCn to transmit Redo logs
By default, ARCn is used to transmit redo logs. However, ARCn can be used only in the highest-performance mode. (For details, referIntroduction to three Protection modes of [Oracle] Data Guard), And use ARCH to transmit Redo logs as follows:

The general process is as follows:
1) Once the log switch is completed for the Primay segment ARC0, ARC1 transfers the newly generated archive log to the Standby end;
2) logs are received by the RFS process on the Standby end. If the standby redo log is configured, logs are recorded to the standby redo log. Wait for the standby redo log to perform log switch to form an archive log, and then archive logs are applied for recovery; if the standby redo log is not configured, after the RFS process receives the log, it is placed in the archive directory of the standby end, and standby then applies the archive log for recovery.

2. Use LGWR to transmit Redo logs
The LGWR process is very different from the ARCn process. The most obvious difference is that it does not need to wait for the Primary to complete the log switch and then transmit the data, as shown below:

The process is roughly as follows:

1) Once Primary has a Redo log, LGWR will trigger the LNSn process to transmit Redo logs only Standby redo logs. Note: LGWR cannot directly transmit the logs because the entire database instance has only one LGWR, to ensure that its main performance is not affected, it cannot directly transmit it)

2) You can select sync or async for the network transmission mode. sync means that when the Primary is submitted, it must wait until the Redo is successfully transferred to Standby before returning. Therefore, if sync is set, we recommend that you set the NET_TIMEOUT parameter at the same time. If no response is returned during timeout, an error is returned. Async means that whether the submission of Primary is successful has nothing to do with whether the log is successfully transferred. This minimizes the impact on the performance of Primary.

3) The RFS process at the Standby end writes Redo to the Standby redo log. If real-time applications are enabled, the redo log is applied to the Standby database. If real-time applications are not enabled, wait until the Standby redo log is archived and then applied to the Standby database.