What is happening in MySQL space?
  * comercial & non-commercial development

Clustrix - mysql cluster appliance
New Relic - monitoring / troubleshooting service
Virident - PCIe flash storage

MySQL 5.1 GA in Nov of 2008
  * Last version done by main mysql team in its entirety, after Sun acquired, but before Oracle

MySQL 5.5 came out in december of 2010
  * 2 years after 5.1, 3 years between 5.0 and 5.1
  * 5.5.8 is first stable release, less releases to get stable
  * mainly cleanup and performance focused
    * better plugins interface
    * PERFORMANCE_SCHEMA biggest user-facing change (for DBAs, analyzing performance)

MariaDB
  * can plugin different storage engines
  * Is a drop-in replacement for MySQL, from some of the original team
  * MariaDB 5.1 and 5.2 based on MySQL 5.1
  * Has virtual columns, define derived values, presented as a column
  * v5.3 has a lot of focus on optimizer

Drizzle
  * Reinventing MySQL -- remove everything that's "wrong"
  * currently in beta, RC soon
  * changes to app required to run on drizzle
  * active work occuring, high commit rate

Percona Server & XtraDB
  * Mysql-code based product, focused on performance, compatibility, and diagnostics
  * a lot of focus on flash storage, and running in the cloud
  * 5.5 beta now available

Community is still producing patches
  * Mark Callaghan, first at Google, now facebook, is leading a lot of community charge, pushing oracle team to solve issues
  * Small patches by independent authors
  * Many things can be plugins now (v5.5)
    * Memcache udfs
    * Handler socket
    * "sqlstats" statement monitor plugin - cool state and monitor plugin

Amazon RDS
  * simplifies running mysql in cloud
    * replication, availability, patching, monitoring
  * familiar api based management (AWS api)

Other performance projects:
  * Schooner MySQL - focused on multicore, SSD
  * TokuDB - storage engine with fast insert
  * ScaleDB - like oracle raq for mysql
  * Clustrix - not based on mysql code, just supports mysql protocol

Data Warehouse - mysql optimizer not strong here
  * requires new optimizer, as well as storage engine
  * most DW solutions aren't row-based, ala mysql
  * InfoBright - high compression, more SQL features
  * InfiniDB by calpont - multi core and multi server
    * early stage, not as much compression, sql features supported
  * Kickfire
    * hw accelearted data flow execution
    * failed

Bridging SQL and noSQL
  * going to see a mix going forward, of some data stored in mysql, other data in nosql
  * HANDLER API in mysql since early days
  * NDB Cluster always had native API
    * ways to access data in mysql DB, without using SQL
    * this is actually popular, as opposed to handler interface
  * PBXT - scalable blog streaming project
  * Handler Socket - new plugin for mysql 5.5, included in percona server 5.5
  * Sphinx full text search engine added support for mysql protocol
    * other nosql products might start supporting mysql protocol?
  * hive interface, sql interface to map/reduce

Xtrabackup
  * innodb hot backup, commercial product popular for awhile
  * can do LVM, but requires integration w/ sysadmin team
  * Xtrabackup is open source implementation of same product
    * command-line compatible for basic features
    * pioneered incremental backup, streaming, parallel backup
    * backup multiple tables at a time
    * done work to limit performance impact on DB, can rate limit speed of backup w/ option
  * zmanada is working on integrating xtrabackup into their backup solution for mysql

Helper tools
  * maatkit
  * open ark kit - nice set of tools, written in python, not as well known
  * aspersa - set of simple tools, for collecting stats from system and mysql, written in shell and awk (highly portable)
    * percona offering free mysql performance review, using this
  * MMM - mysql master/master replication manager, helps with managed failover

Replication
  * row-level replication added in mysql 5.1
    * still untapped potential - could do parallel replication, etc.
  * v5.5 adds semi-synchronous replication
  * percona server has crash safe replication for innodb
    * if slave server crashes, can restart from proper binlog position
  * still single threaded -- tungsten replicator solves this

Hive interface integrates with hadoop
Continuent interested in moving data back/forth
Oracle acquisition has had one positive -- now innodb & mysql under the same roof, so innodb developemnt has picked up the pace
  * pace of innovation is faster
  * not being competitive with oracle db product
  * no intention to kill it
  * long time ago, mysql team had goal to replace oracle, handle enterprise applications like sap
  * conscious effort to make products non competitive, but integrated
    * bi-directional integration between two