Replication
Keep copies of data on several machines - leader-follower, sync versus async, lag, multi-leader and quorums.
Why copy data
Replication means keeping the same data on more than one machine. Three reasons:
- Availability: if one machine dies, another still serves the data.
- Read scaling: spread read traffic over several copies.
- Latency: keep a copy near users in another region.
It does not help with data that is too big for one machine. That is sharding, the next lesson.
The hard part is not copying once. It is keeping copies in agreement while writes keep arriving and machines keep failing.
1 of 9