SQL and Ad Hoc Analytics with MariaDB ColumnStore

MariaDB ColumnStore extends MariaDB Enterprise Server with distributed, columnar storage and a massively parallel processing (MPP) shared nothing architecture, transforming it into a standalone or distributed data warehouse for ad hoc SQL queries and advanced analytics without the need to create indexes.

Columnar

Data is stored in a columnar format for fast analytics and with up to 90% compression, reducing disk I/O and size of data on disk.

Distributed

Data can be distributed across multiple database instances and queried in parallel to increase query performance and support massive datasets.

Cloud-Native

Data can be stored on Amazon S3 compatible object storage, on premises or in the cloud, to benefit from low-cost, unlimited storage.

See What's New in ColumnStore

Customers

Customer Story

Institute for Health Metrics and Evaluation (IHME)

Learn More

Customer Story

How Pixid dropped Oracle and went hybrid with MariaDB

Learn More

real-world analytics use cases

MariaDB ColumnStore supports our customers across various industries and delivers reduced costs, increased performance and better scalability than alternative solutions.

Financial Services

Analyze millions of quotes for trading in tens of milliseconds to estimate price point, profitability, and risk.

MariaDB in Financial Services

Healthcare Analytics

Analyze disparate and diverse data across multiple database instances and increase query performance and support massive datasets.

IHME Case Study

Retail

Analyze customer behavior for retail use cases such as cross-sell/up-sell, customer loyalty and churn.

Manufacturing

Analyze ERP and IoT data to predict points of component and systems failure thereby increasing uptime and reducing maintenance costs.

MODERN ANALYTICS CAPABILITIES

Standard SQL

Analyze large amounts of data with standard SQL, including joins with other tables (row or columnar).

Kafka and Spark connectors

Use Kafka and Spark connectors to ingest streaming data and publish machine learning results for interactive, ad hoc analysis.

Direct data loading

CLI import tools as well as C, Java and Python connectors bypass the SQL layer to load data directly to storage from anywhere.

CORE COMPONENTS For Analytics

MariaDB ColumnStore: A specialized storage engine for scalable, high-performance analytics without the need for complex schemas and indexes – distributed, columnar storage with parallel query processing.
MariaDB MaxScale database proxy: Provides enterprise high availability, scalability, security and integration services, simplifying application development and DB administration.
MariaDB Enterprise Server: A modern, general-purpose database engineered for performance, reliability, security and extensibility with a modular architecture.
Bulk data adapters: C++, Java and Python libraries for enabling developers and data scientists to collect and write data to storage on demand, bypassing the SQL interface for higher throughput.
Streaming data adapters: Apache Kafka and MariaDB MaxScale services for importing data from MariaDB or Apache Kafka in near real time – automatically, transparently and continuously.

ColumnStore Diagram