As organizations move beyond initial testing of Hadoop, enterprise level requirements are increasingly important. Enterprise readiness not only constitutes critical elements such as high availability, security and data protection but also multitenancy capabilities that ensure the ability to effectively share a cluster across users and solutions.
MapR created an enterprise-grade distribution for Hadoop by making significant enhancements to improve Hadoop integration, operation, scalability and reliability. MapR rearchitected the storage services layer to overcome the many limitations of the Hadoop Distributed File System (HDFS), resulting in a file system that provides full random read-write support for multiple concurrent streams.
In rewriting the storage layer, MapR created a binary-compatible HDFS API so all MapReduce programs execute with no changes. MapR also provided support for direct access to data via the industry standard Network File System (NFS) protocol, meaning that any application or tool can access the MapR cluster directly in real time. The fully read-write storage layer, in conjunction with NFS, enables real-time work flows that are more nimble than the batch-oriented processes of the past.
MapR also invested in another important area of innovation, High Availability for all of the critical components, including the NameNode, JobTracker and TaskTrackers. These enhancements provide automated stateful failover against multiple failures, making MapR the only distribution for Hadoop to provide full automated fault tolerance.
With innovations such as these, MapR has created the only truly highly available distribution for Hadoop, and is uniquely positioned for mission-critical, SLA driven, growing enterprise environments. Visit www.mapr.com for more information.