MyISAM

This is a light and nontransactional engine. It has good performance and a small data footprint. MyISAM has good performances for small database access.

InnoDB

InnoDB gets very high performance when the database size is less or equal to the RAM size to be used as cache. It's unfortunately not as efficient as MyISAM if there is not enough RAM memory.

BLACKHOLE

The BLACKHOLE engines accept data but they are immediately dropped and a zero result is returned by the engine. This engine is generally used for complex replication filtering on very high load databases.

CSV

The CVS engine is able to read and write CSV (comma-separated-values) format files.

MEMORY

The MEMORY engine stores data in memory to perform very fast queries. Generally used for read-only cache or temporary access, it's the fastest engine but with some limitations, such as no support for blob or text columns. As a result, it's a nonpersistent engine because all data is in the RAM.

ARCHIVE

The ARCHIVE engine is a good solution for minimal disk space occupation with small footprint. It compresses rows with the zlib algorithm. It's one of the slowest solutions (no index + compression) but it's perfect to store a huge amount of data without specific performance requirements.

MERGE

The MERGE engine is a collection of identical MyISAM tables that can be used as one table.

Aria

>= 5.1

This engine is a crash-safe alternative to MyISAM. It's able to be a transactional and nontransactional storage.

XtraDB

>= 5.1

The Percona XtraDB engine is a drop-in replacement for InnoDB. It's more scalable with many cores and also gets highest performances and more metrics.

PBXT

< 5.5

The PrimeBase XT (PBXT) engine is designed for a high concurrency environment. It's unfortunately not maintained anymore.

FederatedX

>= 5.1

This engine is a drop-in replacement of the FEDERATED engine. It uses libmysql to talk to an RDBMS. The idea is to use other RDBMS as data sources.

SphinxSE

>= 5.2

The Sphinx Search Engine (SphinxSE) is a built-in client to talk directly to the searchd daemon and run search queries. It doesn't store data at all.

IBMDB2I

< 5.5

This engine is able to store its data in a DB2 table running on IBM.

TokuDB

>= 5.5

This is a highly scalable engine with indexing-based query acceleration, no slave lag performance, unparalleled compression, and hot schema modification. It has better performances compared to XtraDB, when you do not have enough RAM. In that case, performances are quite similar to MyISAM.

Cassandra

>= 10

This engine allows direct access to a Cassandra cluster from MariaDB.

Connect

>= 10

This engine permits access to local or remote data when defining tables based on different data types.

Sequence

>= 10

This engine permits the creation of ascending or descending number sequences using a starting and ending value and increment.

Spider

>= 10

This is a built-in sharding features engine that supports xa transactions, partitioning, and allows table links to the table on a remote server.

HandlerSocket

>=5.3

This is a NoSQL plugin. It doesn't support SQL queries but supports CRUD (Create/Update/Delete) operations on tables. It accepts direct TCP connections.