How does MapR manage to provide customers with the latest open source projects ?
MapR has followed a decoupled release model since 2013. MapR Platform has two types of releases: core releases and ecosystem. The ecosystem releases happen every quarter, allowing MapR to take the latest and greatest innovations, test them, and rapidly make them available to customers. The core releases happen less frequently throughout the year to ensure a solid foundation on which to run ecosystem releases.
- Core Release (About 4 releases per year (2 feature, 2 maintenance)) consists of
- MapR Converged Data Platform
- Other core Hadoop components
- Hadoop commons framework
- MapReduce v1 (JobTracker/TaskTracker)
- MapReduce v2
- MapR Ecosystem Packs (MEP) (About 4 per year, roughly once a quarter) consist of one or more of the following
Which versions does MapR support? How do I know if a particular version of one open source project interoperates with a particular version of another open source project?
MapR supports multiple versions of these components. To get the gory details please visit MapR Expansion Pack (MEP)
What about security? Can I run all these projects securely?
When it comes to security, there are multiple options available to match your particular constraints and requirements.
- Security Support Matrix - Latest Documentation - maprdocs.mapr.com
- Explore Security & Data Governance Resources
Is there a difference between the version that is bundled by default and what can be supported by changing the configuration?
Yes. Say for instance the Oozie package bundles some Hive, Pig, and Spark jars. However, if your configuration requires a different version, you can setup a different version. For each project's interoperability with other projects, see the Ecosystem Support Matrix - Latest Documentation - maprdocs.mapr.com
Can I get the latest and greatest open source project version even if its cutting edge and may not be fully interoperable?
MapR provides development previews of the latest versions of Apache Drill and Apache Spark. For instance, Apache Spark 1.6 development preview steps are available at Spark 2.0.0 Developer Preview - Latest Documentation - maprdocs.mapr.com
Note that development previews are not supported for production use and are provided 'as-is' with community/forum support to enable developers to experience the latest technologies.
Can I run related open source projects even though MapR does not provide a supported version of them?
Yes. Customers of MapR run other related projects like Apache Apex, Apache Flink, Apache Zeppelin, Apache Kafka, and several others.