MySQL5.5 New Features
Standard Storage Engine Replacement
MyISAM has been replaced with InnoDB in MySQL 5.5.
InnoDB offers a transaction function based on the efficient ACID. It has a recovery feature for error occurrences and a reference consistency feature which is compatible with transaction applications. The default storage engine has been changed but MyISAM and other types are usable. Due to this change, performance has been improved. Diagram 1 are results of a benchmark test run with SysBench on a Linux. Compared with 5.1 even when connected users increased, the number of transactions per second did not drop. Looking at readings, there is a 200% increase from 5.1.
Also on a benchmark test with SysBench on Windows, the performance and scalability of MySQL 5.5 surpasses 5.1 by 540%.
As a replication of the prepared master and slave, data consistency is managed. When data is being updated, after update of the slave log has been confirmed, it is set up so the master will commit a transaction. With this function, data consistency can now be maintained.
Heartbeat is a message sent at fixed intervals from master to slave. On the slave side an automatic check for a connection to a master is carried out by message checker. In a case where the check result did not receive the message, an error in connection will occur and the fact that feature ② is not operating normally will be known. This is a feature to guarantee feature 2.
Monitoring of processes spent on server events, time, and resources is possible. It is different from the INFORMATION_SCHEMA of traditional recording and analysing of MySQL database metadata. With this feature, collected data and events are stored in a table within the new performance_scheme database. The confirmation of this data uses the SELECT text. Execution performance of a server is monitored and this can assist operation analysis or improvement.
Internal InnoDB management has been strengthened. Special use cases and application loads have been optimized. As features for optimization, index creation has been accelerated and data compression has been optimized. For large objects and variable length lines, efficient storage options and a new INFORMATION_SCHEMA table has been added.
Addition of SQL and Partition options
Stored objects or tables and index partition features that were implemented in previous versions have been improved. As for specific details, the appearance of a new SQL syntax: Signal/Resignal, non-integer columns and dating partitioning, and partitioning of several lines has been realized. From this, developing efficiency will be improved.