# `2-DC` with `Binlog` Replication Deployment Solution

![`smart`](2DCBingLogDeployment_aa_001_20221025.png)
------------------------------------------------------------

Two data centers with `Binlog` synchronization is similar to the traditional primary/secondary plan in `MySQL`. 

The two data centers each deploy a standalone `TiDB` cluster, which we call the source cluster and the sink cluster. Under normal circumstances, all requests are made to the source cluster, and the written data is asynchronously synchronized to and written from the cluster through `Binlog`.

When the data center hosts source cluster fails, the business can switch to the sink cluster. Similar to `MySQL`, some data will be missing in this case. Compared to `MySQL`, the advantage of this solution is `HA` in the data center — when a small number of nodes fail, the service is automatically recovered by re-electing the `leader` without manual intervention.