MapR Technologies, provider of the Apache Hadoop distribution, has open sourced their MapR-DB NoSQL database for unlimited production use.
MapR-DB, part of the MapR Community Edition (formerly M3), is based on Google BigTable and supports a wide column NoSQL data model. It has native integration with Hadoop, supports automatic sharding and can be used for use cases in the areas of data streaming and real-time analytics. Some of the datasets to store in MapR-DB include operational data formats like logs, sensor data, metadata, and clickstreams. The database uses Apache HBase core API to run existing HBase applications along with MapR-DB. Developers can run applications using HBase and MapR-DB in the same cluster.
MapR-DB database also supports other features like the following:
- Row-level ACID transactions
- Strong data consistency
- Integrated Security which includes Kerberos or LDAP integration for authentication and Access Control Expressions for authorization
- Multi-tenancy
- In-memory database option that can be used to store an entire database in memory for additional performance gains
It doesn't support High Availability and Disaster Recover (HA/DR) capabilities.
MapR-DB in the MapR Community Edition is available for download on the company website. There is also the MapR Sandbox, a fully-functional Hadoop cluster running in a virtual machine for developers to test their product without a stand-alone installation or configuration overhead.