Ordered indexes in MySQL Ndb Cluster allow efficient [ordered] range scans on primary or secondary keys.

MySQL Ndb Cluster ordered indexes are implemented using the T-tree structure.

T-trees store fixed-size direct row references, avoiding duplication of indexed columns or Primary keys. The index storage cost is independent of the indexed columns' size.

Index partitions are co-partitioned and co-located with table partitions, allowing efficient atomic index maintenance.

Index partitions are built, searched and modified in parallel.

https://en.m.wikipedia.org/wiki/T-tree

#DistributedSql #MySQL #ndbcluster

MySQL Ndb Cluster partitions, distributes and scales out tables and indexes, enabling parallel scan, search, filter and join within + between SQL queries.

https://messagepassing.blogspot.com/2011/03/data-distribution-in-mysql-cluster.html

https://mikaelronstrom.blogspot.com/2019/12/ndb-parallel-query-part-1.html

#DistributedSQL #mysql #ndbcluster

MySQL Ndb Cluster has built in automatic sharding and synchronous replication for HA + scalability.

Ndb HA has some unusual features that make it more efficient and capable than many comparable systems.

Using a stateless arbitrator allows Ndb to offer partition safe HA with just 2 replicas of data, as opposed to the common baseline of 3, saving resources.

Dynamic quorum reconfiguration allows an s shard, r replica system to survive the loss of all but one replica of each shard, as opposed to only surviving the loss of less than half the shard replicas.

#mysql #HA #ndbcluster #distributedsql

MySQL Ndb Cluster is GPL open source C++. Everyone can browse the source code, download and build it, modify it, fork the whole project, run it on premises, in the cloud or as a service.

https://github.com/mysql/mysql-server/tree/trunk/storage/ndb
https://dev.mysql.com/doc/refman/8.4/en/source-installation.html
https://dev.mysql.com/downloads/cluster/
#mysql #opensource #ndbcluster #ha #distributedsql

MySQL Ndb Cluster is a variant of MySQL and as such supports the MySQL JSON functions, the MySQL JSON type and related functionalities. These mechanisms give the database visibility into stored JSON documents, enabling more efficient storage and processing without tightly constrained schemas.

JSON columns can be stored, retrieved, and queried from MySQL Ndb Cluster in a structured way via SQL or other APIs (MySQL X API, Document store, REST server implementations etc).

JSON data can be selectively indexed using stored generated columns, allowing efficient access or the use of relational constraints.

https://dev.mysql.com/doc/refman/8.4/en/json.html

https://dev.mysql.com/doc/refman/8.4/en/json-functions.html

#mysql #distributedsql #ndbcluster #opensource

MySQL Ndb Cluster is often described as an in-memory or main memory database but this is not the whole story.

Tables in Ndb always support :
- Transactions, parallel query, indexes, FKs etc
- Sync replication between nodes in a cluster
- HA with automatic failover and recovery
- Consistent backup
- Async replication via MySQLD

By default Tables in Ndb also have automatic transactional durability via parallel checkpoints and logs, supporting full system recovery from storage. However, this can be disabled for individual tables if desired for highest performance with lowest resource usage. In the event of a full system recovery from storage, these tables will be empty.

By default, all columns of a table are kept in memory at all times, limiting system capacity to full cluster system memory size. However unindexed columns can be marked for disk storage. Values for these columns are stored in pages of a tablespace, and only cached in memory like most other RDBMS. This allows far more data to be stored while retaining predictable fast access via purely in memory indexes.

#MySQL #ndbcluster #DistributedSQL

MySQL Ndb Cluster 8.4 specific security features include :
- Encrypted backups
- Encrypted storage (Checkpoints, redo logs, undo logs + tablespaces) (TDE)
- mTLS on all internal connections for authentication and encryption (Ndb TLS)

All open source, implemented using industry standard optimised and reviewed OpenSSL libraries.

These features are in addition to the standard MySQL Server Security features ([m]TLS, TDE, Binlog encryption, PAM etc)

https://dev.mysql.com/doc/refman/8.4/en/mysql-cluster-security.html
https://dev.mysql.com/doc/refman/8.4/en/mysql-cluster-backup-using-management-client.html
https://dev.mysql.com/doc/refman/8.4/en/mysql-cluster-tde.html
https://dev.mysql.com/doc/refman/8.4/en/mysql-cluster-tls.html

https://dev.mysql.com/blog-archive/mysql-cluster-8031-transparent-data-encryption/
https://blogs.oracle.com/mysql/introducing-ndb-tls
https://messagepassing.blogspot.com/2020/11/mysql-cluster-backup-and-restore.html

https://dev.mysql.com/doc/refman/8.4/en/security.html

#DistributedSQL #MySQL #NdbCluster #OpenSource

MySQL Ndb Cluster quarterly releases are now available :
9.6.0 https://dev.mysql.com/doc/relnotes/mysql-cluster/9.6/en/
8.4.8 LTS https://dev.mysql.com/doc/relnotes/mysql-cluster/8.4/en/news-8-4-8.html
8.0.45 https://dev.mysql.com/doc/relnotes/mysql-cluster/8.0/en/news-8-0-45.html
7.6.37 https://dev.mysql.com/doc/relnotes/mysql-cluster/7.6/en/news-7-6-37.html

#mysql #ndbcluster #distributedSQL

Good to see that RonDB has a slot in the 2026 FOSDEM databases devroom.

They are doing interesting work extending NDB to cover new and trending use cases that you may have heard of.

https://fosdem.org/2026/schedule/event/real-time_ai_powered_by_rondb/

#DistributedSQL #ndbcluster #fork #opensource