This is part 2 of a three-part review. You can find part one here.
5. Replication
This is the first chapter in the Distributed Data section. Replication means that the same data is stored on multiple machines. Some reasons for replication are: to keep working even if some parts of the system fail (increased availability), to keep data geographically close to the users (reduce latency), and to increase read throughput by serving the same data from many machines.