Re:Hadoop is a distributed computing platform (Score 1)

....

There is a distributed filesystem built on top of hadoop called HDFS.

...

I would say that Hadoop runs on top of HDFS. That is, the Hadoop layer expects to use a distributed filesystem which has some location awareness (allowing processing tasks to be sent to the nodes containing the data) and a level of resilience to individual datanodes failing (there is at least 2-way replication of data assumed in a typical HDFS configuration). The HDFS API expected by Hadoop can be implemented by other filesystems, such as Kosmos File System, mentioned above.

Apos for posting this comment again, but I had the suspicion that posting anonymously was hiding my comment...