[Myhadoop-users] Hadoop for HPC

[Ramkumar C. <ram...@gm...>]

Hello myHadoop Users,

This is Ramkumar from University of Washington. In UW, we have a large HPC
cluster system called HYAK where we submit our jobs. We are currently
exploring all resource available for highly scalable data analytics jobs
(Mostly analytical and few I/O bound operations) Even with the HPC
resources, we are still bottle-necked by I/O speed, hence we are thinking
of building a cluster of HPC resources (cluster with few HPC machines in
the order of 10's).

It during this process, we found your
presentation<http://www.sdsc.edu/us/consulting/myHadoop-SDSC.pdf>
(PDF)
and your paper on the same. We are exciting in exploring this as it sounds
like a very nice solution to the problem that we have at place. The latest
version of myhadoop from SourceForge (myhadoop -0.2a) that we are using
uses Hadoop 0.2 in the backend( which is know is quite older version of
Hadoop) We are interested in knowing if you guys have proceeded to have
this compatible with the latest versions of Hadoop ? Also if you have
suggestions/alternatives for running Hadoop in HPC environments, I would be
really appreciate the help. Thanks in advance!

Regards,

Ramkumar Chokkalingam ,
University of Washington.
LinkedIn <http://www.linkedin.com/in/mynameisram>