willdignazio / hrfs Goto Github PK

		    Hadoop Replicating FileSystem


What is it?
-----------

The Hadoop Distributed Filesystem (HDFS) is a modern distributed filesystem
implementation, built upon the Hadoop distributed architecture platform. By
using the API and facilities within hadoop, HDFS proves to be a fault tolerant
system that is capable of streaming data to and from constituent nodes. However,
there is a catch: HDFS’ use of a single Nameserver for metadata and filesystem
operation management becomes a bottleneck under the wrong workload.

The Hadoop Replicating FileSystem (HRFS/hrfs) aims to solve this problem by
distributing file data and metadata replicants to participating data nodes. In
doing so, it hrfs increases the availability of the data, and provide higher
throughput for more desired content. It incorporates a modular design that
facilitates fault tolerance by removing single points of failure within the
system as a whole.

Further, hrfs aims maintain a dynamic level of replicas for file content, such
that replicas from alternative hrfs nodes may service requests when the one node
is under duress. With increased availability comes the need to effectively
gather it in a large cluster, this is achieved through late binding of the
request for file data; When a request for a file occurs, the the least loaded
hrfs node will shall be the one to actually serve the content.

The latest Version
------------------

The latest version and details can be found on GitHub,
https://github.com/WillDignazio/hrfs/.

Documentation
-------------

The documentation for Hrfs can be generated with maven, and can be invoked with:

    $ mvn javadoc:javadoc
    $ cd target/site/apidocs
    $ python -m SimpleHTTPServer 8080

This will generate the documentation for hrfs, and render it as a browsable
source on http://localhost:8080/.

Further documentation can be found in docs/, which contains various design and
configuration information.

Licensing
---------

Please see the file called LICENSE.

Contacts
--------

	* For all issues and bug fix submissions, please use the GitHub
          interface at https://github.com/WillDignazio/hrfs/.
	* For questions, comments, and other project related requests, email
          wdignazio(at)gmail(dot)com or mxkvcs(at)rit(dot)edu.