MySQL 5.6 improves developer agility with subquery optimizations, online Data Definition Language (DDL) operations, NoSQL access to InnoDB, new instrumentation in Performance Schema and better condition handling.
Highlights of MySQL 5.6 include:
Better Query Execution Times and Diagnostics provided through an enhanced MySQL Optimizer diagnostics Subquery Optimizations
Simplify query development by optimizing subqueries prior to execution. New efficiencies in how result sets are selected, sorted and returned deliver substantial improvement in query execution times. The Addition of Index Condition Pushdown (ICP) and Batch Key Access (BKA): can improve selected query throughput by up to 280x.
Enhanced Optimizer Diagnostics: with EXPLAIN for INSERT, UPDATE, and DELETE operations. EXPLAIN plan output in JSON format delivers more precise optimizer metrics and better readability, and Optimizer Traces enables to track the optimizer decision-making process.
Greater Performance Throughput and Application Availability with an improved InnoDB storage engine
Up to 230 Percent Improvement in Transactional and Read Only Throughput: InnoDB has been re-factored to minimize legacy threading, flushing, and purge mutex contentions and bottlenecks, enabling better concurrency on heavily loaded OLTP systems, and resulting in significantly improved throughput for both transactional and read only workloads.
Online DDL operations enable DBAs to add indexes and perform table alterations while the application remains available for updates.
Full-Text Search with InnoDB allows developers to build FULLTEXT indexes on InnoDB tables to represent text-based content and speed up application searches for words and phrases. Flexible NoSQL access to InnoDB provides simple, key-value lookup of InnoDB data via the familiar Memcached API. Users get the “best of both worlds,” combining key-value operations and complex SQL queries in the same database.
Improved Scale-Out and High Availability
With new features in MySQL replication including Self-Healing Replication Clusters. The addition of Global Transaction Identifiers and Utilities make it simple to automatically detect and recover from failures.
Crash-Safe Replication enables the binary log and slaves to automatically recover correct positions in the replication stream in case of a crash, and resume replication without administrator intervention. Checksums maintain data integrity across the cluster by automatically detecting and alerting on errors.
Up to 5 times faster replication through Multi-Threaded Slaves, Binlog Group Commit and Optimized Row-Based Replication enable users to maximize the performance and efficiency of replication as they scale-out their workloads across commodity systems. Time-delayed Replication provides protection against operational errors made on the master, for example accidentally dropping tables.
New instrumentation enables users to better monitor most resource intensive queries, objects, users and applications. New summaries with aggregated statistics grouped by query, thread, user, host and object are also available. The enhancements allow for easier default configuration with less than five percent overhead.