MySQL Cluster

2016: MySQL Cluster 7.5.0

On February 6, 2016 Oracle provided release of MySQL Cluster 7.5.0^[1].

The packet helps to organize the distributed storages and highly reliable configurations which will provide the level of availability of service ~ 99.999% when providing requirements of ACID to accomplishment of transactions (atomicity, coordination, isolation, durability).

The project code is distributed under the license GPL and is available to free loading. The release is remarkable transition to use of a branch of MySQL 5.7 and updating of the NDB engine.

MySQL Cluster provides means for creation of a distributed network of the servers replicated in the multi-master mode guaranteeing lack of a uniform point of failure. A system provides horizontal scaling - accumulation of power of a cluster is made due to connection of new nodes and use of the technology of automatic sharding (distribution of a data set by servers on the basis of a certain key). For solving of tasks of a real-time mode the storage for data processing in RAM (In-Memory) is offered.

SQL and NoSQL API, including interfaces are available to C ++, to Java, http, Memcached and JavaScript/Node.js.

2015

There was a release of MySQL Cluster 7.4.4

On February 26, 2015 MySQL development team as a part of Oracle announced readiness of MySQL Cluster 7.4 for production loads^[2].

According to developers, in it there were no essentially new features, but acting considerably are improved.

Screenshot of diagrams of performance (reading/record transactions / c) different versions of MySQL Cluster, 2014

Advantages

• The maximum number of nodes of a cluster remained without changes, but in the biggest configurations significantly (up to 50%) performance is increased. Function of their parallelization between cluster nodes with the subsequent consolidation of results was added to an opportunity to execute difficult connective operations (join) that considerably improved scalability. In similar tasks special importance is purchased by the scanning speed of tables which in the specific version is also increased.

• Management tools and administrations are improved. The possibility of transparent adding of a new node operating earlier in a cluster demanded functions of control of distribution of data between nodes (the model without separation of system resources is admitted to MySQL Cluster). As separate tables DB can be stored in memory (in-memory) and be distributed on RAM of different nodes of a cluster, transactions of removal in them should be followed by redistribution of data on nodes and release of the OZU resources for the subsequent use. These functions are implemented in version 7.4.

• MySQL Cluster is designed so that in it there is no deniability at failure of one of nodes. In particular, for this purpose completion of transaction is fixed only after synchronous replication of all changes on different nodes of a cluster. At the same time also asynchronous replication which in MySQL Cluster is called geographical and is applied to ensuring disaster tolerance by reservation of clusters in geographically remote data centers is supported. In version 7.4 similar replication gained further development that, according to creators, allowed to narrow temporary windows for first technical support. It is important for the companies with the business which is geographically distributed on a set of time zones (telecommunication operators, on-line retailers).

2014

MySQL Cluster 7.4

On November 6, 2014 the Oracle company announced release of the update - the DMR version of MySQL Cluster 7.4.

As a part of new MySQL Cluster 7.4 a number of improvements for the purpose of improvement of dynamism, scalability, duration of trouble-free operation is offered:

• The replication in the mode "active-active" (active-active replication) allowing to replicate data between the distributed clusters with automatic detection and permission of conflict situations.

• Performance improvements – at the expense of a possibility of use of bigger number of cores in each node for "large-scale" horizontal scaling of capacity.

• Operational improvements, such as improvement of the reporting and acceleration of transactions of service. The key functions which are in a development stage, "early access" to which is open for testing and discussion by community MySQL:

• Multi-source Replication — consolidates data from several master-servers on one or more slave-server.

• MySQL Group Replication — simplifies ensuring high availability, allowing any server to accept records (this), and increases transparency of applications due to elimination of need of coordinating between the application and routing servers of transactions.

2013

MySQL Cluster 7.3

On June 19, 2013 the Oracle corporation announced release of the new version of MySQL Cluster 7.3.

The new version of the software package for deployment of cluster configurations on the basis of MySQL DBMS simplifies to developers of a possibility of expansion of their applications the database of real time with high availability, scalability and fault tolerance.

The software package helps to create the distributed storages and highly reliable configurations which are capable to provide availability of service at the level of 99.999% when providing requirements of ACID ^[3] to accomplishment of transactions. The code is distributed under the license GPL and is available to free loading.

Existence of NoSQL JavaScript Connector library for the node.js platform allows to create the services intended for deployment on the cluster configurations consisting of standard hardware with the minimum efforts to development and management.

New opportunities

• the built-in support of foreign keys (Foreign Keys), for control of connectivity and integrity of data in the tables distributed on different nodes of a cluster, including which are in different data centers;

• automatic installer on the basis of the browser by means of which it is possible to start the solution on the basis of MySQL Cluster in a few minutes and to optimum configure a configuration, depending on required type of the tasks solved by a cluster;

• the optimized scalability in flows of processing of connections which in addition help to provide for end users compliance to requirements to the database of high availability imposed by new generation cloud, communication and web services.

The layer of SQL is based on the latest version of MySQL 5.6 now that allows developers and system administrators to use such features of the version as the increased capacity and reliability of replication. Using the new version of MySQL 5.6, developers can combine mechanisms of data storage InnoDB and MySQL Cluster in one application.

The scalability for flows of processing of connections (Connection Thread Scalability) provides capacity growth by 1.5 — 7.5 times, counting on one connection with nodes of data of a cluster MySQL Custer, increasing the general capacity and scalability of a cluster. Improvement is reached due to splitting internal blocking and reduction of the size of critical sections in a processing code of connections.

2012

MySQL Cluster 7.2

The Oracle corporation announced at the beginning of 2012 a release of the new version of a relational database of MySQL Cluster 7.2 intended for providing 99.999% of availability, high scalability of record and midget response time. The new version of MySQL Cluster supports both language of requests of SQL, and model of access of NoSQL via the new Memcached API interface, at the same time provides performance improvement at accomplishment of advanced queries and scalability for DPCs with several sources (multi-data centers).

According to developers, MySQL Cluster 7.2 allows to unroll the distributed, high-scalable databases with both interfaces — SQL and NoSQL, with a possibility of accomplishment of advanced queries or multitabular transactions according to requirements of the ACID standard. Users can execute both difficult, and simple requests — the key/value type (key-value) — for the same data sets in the same database.

"The performance improvement and flexibility provided to MySQL Cluster 7.2 provides to users the reliable platform for crucial web applications integrating advanced technologies of SQL and NoSQL for risk reduction, costs and simplification of a system" — Tomas Ulin, the vice president of Oracle for development of MySQL said.

According to Oracle, adaptive localization of a request provides increase in productivity when processing advanced queries by 70 times. As a result, even more applications, such as analytical applications of real time working with operational data can take advantage of MySQL Cluster.

Among other improvements of MySQL Cluster 7.2 it is possible to note expanded scalability. Multisite clusters (multi-site clusters) allow to place separate nodes with data in different data processing centers; at the same time databases are automatically distributed between these nodes. Synchronous replication maintains integrity and data consistency between "websites", together with opportunities of the fast automated bypass of failures and recovery. Expanded replication in the mode "active-active" (active/active replication) simplifies detection and the conflict resolution between several active clusters, exempting developers from need to support a column of time marks in applications, explained in Oracle.

Other feature of MySQL Cluster of version 7.2 is simplification of use and administration. Shared tables of user rights consolidate the previous distributed tables in cluster nodes with data, making them available from all MySQL servers. Thanks to a new feature administrators do not need to set and manage the user rights on each SQL node of a cluster now.

MySQL Cluster 7.2 is also certified for work with the environment of virtualization Oracle VM. Possibilities of flexible scaling on demand and selfrecoveries in combination with support of Oracle VM do MySQL Cluster by the optimal choice for the solutions unrolled in environments of cloud computing claim in Oracle. Along with Oracle VM, MySQL Cluster DBMS it is also certified for work with the Oracle Linux and Oracle Solaris operating systems.

Notes