RonDB
rondb.com
- Free plan
- Linux
- API
- Self-hosted
What RonDB is
RonDB is an open-source distributed storage engine for real-time applications and a fork of MySQL NDB Cluster. Its shared-nothing design supports up to 1 PiB of storage, with online horizontal and vertical scaling for storage and MySQL-server compute. Applications can access it through the MySQL Server API for SQL-compatible queries, including parallel queries. Through MySQL-server support, it provides ACID transactions, while in-memory data is persisted using REDO logs and checkpoints. The project describes its priorities as low latency, high availability, high throughput, and scalable storage. It also supports AI application storage integrated with Spark and Kafka; Rondis implements a Redis server using RonDB storage and the Redis wire protocol. Deployment options include local testing, Docker, managed cloud use with your own AWS, Azure, or GCP credentials, and on-premises deployment by inquiry. Replication can be synchronous active-active across availability zones or asynchronous across regions with conflict resolution. RonDB is used as the online feature-store database for low-latency, high-availability machine-learning applications in Hopsworks. It is free and supports Linux, API, and self-hosted use.
Who it suits
RonDB suits teams building real-time applications that need distributed storage, SQL access, or an online feature store for machine-learning applications. It may also fit teams using Spark, Kafka, or Redis-compatible access patterns.
Good to know, and what to watch
- Supports up to 1 PiB of storage
- Online horizontal and vertical scaling
- SQL access through the MySQL Server API
- Supports cross-zone and cross-region replication
- Free open-source software
- Uses read-committed isolation
- Does not support range locks
- On-premises deployment is by inquiry
Plans
Every plan and priceServerless RonDB via Hopsworks
Open source
The verdict
RonDB combines distributed storage with SQL access, scaling, and replication options for real-time workloads. Its read-committed isolation and lack of range locks are important transaction constraints to consider.
What the maker publishes
- AI integrations
- RonDB supports AI application storage integrated with Spark and Kafka. source
- Deployment options
- RonDB can be tested locally, run in Docker, managed in AWS, Azure or GCP using own cloud credentials, or deployed on-premises by inquiry. source
- Encryption at rest
- RonDB supports encryption of on-disk data including local checkpoints, tablespaces, REDO logs and UNDO logs, but not backups. source
- Maker support
- RonDB is brought to you by the RonDB team at Hopsworks AB, which provides dedicated support for high-performance online feature-store requirements. source
- Performance model
- RonDB targets low latency, high availability, high throughput and scalable storage, which it calls LATS. source
- Primary use case
- RonDB is used as the online feature-store database for low-latency, high-availability machine-learning applications in Hopsworks. source
- Redis compatibility
- Rondis implements a Redis server using RonDB storage and the Redis wire protocol. source
- Replication
- RonDB supports synchronous cross-availability-zone active-active replication with 1 to 3 replicas and asynchronous cross-region replication across clusters with conflict resolution. source
- REST security
- The RonDB REST API supports Hopsworks API-key authentication and recommends TLS when API keys are used. source
- Scaling
- RonDB supports online horizontal and vertical scaling for storage and MySQL-server compute. source
- SQL access
- Applications can access RonDB through the MySQL Server API with SQL compatibility and parallel query support. source
- Storage scale
- RonDB supports up to 1 PiB of storage. source
- Transactional limit
- RonDB uses read-committed isolation and does not support range locks. source
- Transactions
- RonDB is ACID compliant through MySQL-server support, with in-memory data persisted using REDO logs and checkpoints. source
- What it does
- RonDB is a distributed, shared-nothing storage engine and an open-source fork of MySQL NDB Cluster for real-time applications. source
Checked against the maker's pages on 1 Oct 2026.