In this chapter, we discussed how to spread some files over multiple disks to obtain better performance when the I/O operations are the system's bottleneck. Both table files and logs can be moved to different storage devices. Particularly, moving some InnoDB tables and logs out of the system tablespace can be very beneficial.

We also discussed how to distribute data over multiple servers to balance the workload. MariaDB is distributed with three storage engines that provide these functionalities:

The FEDERATEDX storage engine is designed to access a single remote table. The CONNECT storage engine is used to access external data in a wide variety of formats, and among other data sources, it supports remote servers. The SPIDER storage engine is specifically designed to implement clusters of tables using MariaDB storage engine's API. In all these cases, the communication between the local server and remote servers is transparent for the user, who can query a FEDERATEDX, CONNECT, or SPIDER table just like any other table.

In the next chapter, we will discuss how to implement a cluster of MariaDB databases using Galera.