Like the FEDERATEDX storage engine and the CONNECT MYSQL table type, SPIDER allows remote access to tables as if they were on the local server. However, the SPIDER storage engine has been specifically designed for data sharding. Its main function is to access data from a multitude of servers by querying a single local table.

Data sharding is implemented in SPIDER using table partitioning. If a SPIDER table is partitioned, each partition can be linked to a different remote table. The SPIDER storage engine is suitable with the RANGE and LIST partitioning methods, including RANGE COLUMNS and LIST COLUMNS.

The SPIDER storage engine supports both regular SQL transactions and XA transactions, if the remote tables support them too.

The SPIDER storage engine has been originally designed for MySQL. The version distributed along with MariaDB is slightly modified to take advantage of the MariaDB-specific features.

The original author of SPIDER is Kentoku Shiba. The project's official...