MariaDB Platform X3 combines transaction processing and analytics
Fusing relational and column-store databases into one, with automatic data sync between workload types
22 January 2019 | 0
MariaDB Corporation, developers of the MariaDB open-source fork of MySQL, have announced a new open source database — a fusion of two of its existing products — that processes both transactional and analytical workloads on the same dataset.
“One database, any workload” is how the company is pitching MariaDB Platform X3. The new offering combines two products formerly marketed separately, MariaDB Platform For Transactions and MariaDB Platform For Analytics, into a single, unified solution that incorporates row storage for transaction processing and columnar storage for analytics.
With MariaDB Platform X3, an organisation may use a single database both for conventional customer-facing workloads (transactional, or OLTP) and internal business-intelligence workloads (analytical, or OLAP). The same data is available for either kind of work and is kept automatically in sync between the two sides.
MariaDB Platform is priced at a flat per-node cost, regardless of whether nodes are OLTP or OLAP. This allows for more flexible deployments, where the number of nodes in a given deployment can be moved freely between OLTP and OLAP workloads as demand changes.
In a blog post announcing the release, the company outlined some technical details. The product contains two separate instances of MariaDB server — one for transactional, one for analytical work. Data is stored on the transactional side, with changes and updates synchronised to the analytical side, which uses MariaDB’s ColumnStore engine to handle data for real-time analysis.
Both halves of the new product retain their distinctive features. MariaDB Platform For Transactions provides enterprise features like Oracle compatibility by way of data types and stored procedures, and dynamic data masking for sensitive workloads. MariaDB Platform For Analytics already supported connections for many industry-standard analysis tools, and these are available here as well: Apache Kafka, Apache Spark, Pentaho, and clients for C, Java, and Python. Support for Python allows you to use Jupyter/IPython notebooks with MariaDB Platform For Analytics as a data source.
MariaDB Platform X3 is available both on-premises and via the newly-announced MariaDB Managed Service offering. For on-premises deployments, MariaDB offers conventional binaries and a host of container-based options: simple Docker containers, Docker Compose sandboxes, and Helm charts and scripts for Kubernetes.
IDG News Service