Tokutek Brings Replication Performance to MySQL

Tokutek Inc., in high-performance and agile database storage engines, announced TokuDB v6.0 with high performance replication for MySQL and MariaDB.

It eliminates the common and persistent problem of ‘slave lag’ in which a replication server is unable to keep up with the query load borne by the master server. It solves this by offering high ingestion rates at the slave with best in class compression.

Baron Schwartz, author of the recently updated book, High Performance MySQL: Optimization, Backups, Replication, and More, notes that "Writes are usually the bottleneck in MySQL replication. The single-threaded nature of MySQL replication means it’s relatively inefficient on the slave; even a fast slave with lots of disks, CPUs, and memory can easily fall behind a master."

Until now, organizations typically addressed slave lag by band-aids such as adding more slaves, buying hardware, or sharding the database. Each approach is either expensive, requires one to give up on full replication, or leads to time consuming application re-architecting.

"Replication performance has been a persistent problem for MySQL, and in recent years in particular has become a focal point for innovation in the MySQL ecosystem," commented Matt Aslett, research manager, data management and analytics, 451 Research. "The high data ingestion rates claimed by Tokutek would appear to be a natural fit to helping solve the problem of slave lag in particular."

"Prior to TokuDB, we were using InnoDB for storing all data," said Bastian Axter, director of operations at Evidenzia GmbH & Co. KG. Evidenzia is one of the leading partners of the software, movie, and music industry when it comes to tracing copyright infringements and illegal file sharing activities in peer-to-peer networks. "We found that as the tables grew bigger, the selects were becoming slower, taking as much as an hour or more, and the disk I/O was growing higher." Evidenzia considered storing all the data on SSD, but it was much too expensive. "With TokuDB, the overall impact on disk I/O was impressive, dropping from near 80%-100% down to 5%-10%. We save over 70% on storage," concluded Axter. "We expect TokuDB v6.0 to provide us with additional headroom and capabilities as we continue to expand the service."

With TokuDB v6.0, MySQL users realize:

• Read Scalability (Load balancing) – Replication to several slaves helps distribute read queries across several servers. For read intensive applications, TokuDB v6.0 ensures that analysis and business reports on slaves act on fresh data.
• Disaster Recovery – It enables quick recovery after backup snapshots and ensures up-to-date recovery points after a failure.
• Write Optimized Compression – TokuDB led on compression performance. With version 6.0 it is more aggressive, with compression levels up to 25x.

TokuDB v6.0 also provides better in-memory and point-query performance, two-phase commit for XA, support for MySQL v5.5 (and v5.1) as well as MariaDB v5.2, and improved checkpoint variability.