# `3-DC` Deployment Solution
+ All the replicas are distributed among `3 DCs`
  + In case one `DC` is down, the other 2 `DCs` will initiate `leader` election and resume the service without lossing any data
+ Performance greatly depends on the network latency
  + For writes, all the data has to be replicated to at least 2 `DCs`
  + The read performance will suffer if the `leader` is not in the same `DC` as the `TiDB` node receives the read request
  + Each `TiDB` transaction needs to obtain `timestamp` from the `PD leader`
    + If `TiDB` server and `PD leader` are not in the same `DC`, the performance of the transactions will be impacted by network latency because each transaction with write request will have to get `TSO (Timestamp Oracle)` twice
------------------------------------------------------------
Intra city (<50km), dark fiber, latency < 2ms
Inter city (<400km), MSTP (Multi-Service Transfer Platform), latency < 10ms
Inter city (<1000km), MSTP, latency < 20ms