Database Reference
In-Depth Information
MapR also takes a different approach to storage. Instead of using direct
attached storage in the data nodes, MapR uses mounted network file
storage, which they call Direct Access NFS. The storage provided uses
MapR's file system, which is fully POSIX compliant.
MapR is available both within Amazon's Elastic MapReduce Service and
within Google's Cloud Platform. MapR also offers a free distribution called
M3. However, it is not available in Azure or on Windows and is missing
some of the high-availability (HA) features. For those goodies, you have to
pay to get either the M5 or M7 versions.
Cloudera CDH
Cloudera, whose chief architect is Doug Cutting, offers an open source
distribution called Cloudera Distribution Including Apache Hadoop (CDH).
Like MapR, Cloudera has invested heavily in some proprietary extensions
to Hadoop for their Enterprise distribution. Cloudera, however, also has
an additional release, Cloudera Standard, which combines CDH with their
own cluster management tool: Cloudera Manager. Cloudera Manager is
proprietary, but it is a free download. As far as competition goes, this puts
Cloudera Standard firmly up against Hortonworks's HDP distribution,
which includes Ambari for its cluster management.
Cloudera'sbig-ticketitemisImpala.Impalaisareal-time,massivelyparallel
processing (MPP) query engine that runs natively on Hadoop. This enables
users to issue SQL queries against data stored in HDFS and Apache HBase
without having to first move the data into another platform.
Is HDInsight a Distribution?
In a word, no. HDInsight is a product that has been built on top of the
Hortonworks HDP distribution (specifically the HDP distribution for
Windows). At the time of this writing, HDP 1.3 is the currently available
version.
Core Hadoop Ecosystem
Some projects in the world of Hadoop are simply more important than
others. Projects like HDFS, the Hadoop Distributed File System, are
Search WWH ::




Custom Search