DEPENDENCIES

DEPENDENCIES
------------

Zohmg depends on a number of software packages and Python modules.

The installation procedure described in INSTALL will take care of
these dependencies for you, except for patched versions of 
Hadoop and HBase. Please see scripts/install_hadoop_hbase.sh 
for an assisted installation of Hadoop and HBase.


The following software is required:
 - Apache Hadoop 0.20 with patches HADOOP-1722 and HADOOP-5450.
 - Apache HBase 0.20
 - Dumbo >=0.21.17

These python modules come bundled:
 - PyYAML
 - Python Paste
 - Python simplejson



# Dumbo (>=0.21.17)

    Dumbo is a convenient Python API for writing MapReduce programs.

    more:
     + http://dumbotics.com/
     + http://klbostee.github.com/dumbo/
     + http://wiki.github.com/klbostee/dumbo.

# Apache HBase (0.20)

    HBase is a column-based datastore inspired by BigTable.

    more:
     + http://hbase.org/
     + http://labs.google.com/papers/bigtable.html

# Apache Hadoop (0.20)

    Hadoop is a MapReduce framework.

    Additionally Hadoop source has to be patched with HADOOP-1722 and
    HADOOP-5450:
     + https://issues.apache.org/jira/browse/HADOOP-1722
     + https://issues.apache.org/jira/browse/HADOOP-5450

    more:
     + http://hadoop.org/
     + http://en.wikipedia.org/wiki/Hadoop
     + http://labs.google.com/papers/mapreduce.html
     + http://dumbotics.com/2009/02/24/hadoop-1722-and-typed-bytes/