[openstack-dev] Mahout-as-a-service [sahara]

Matthew Farrellee matt at redhat.com
Thu May 29 11:50:22 UTC 2014


On 05/28/2014 12:37 PM, Dat Tran wrote:
> Hi everyone,
>
> I have a idea for new project: Mahout-as-a-service.
> Main idea of this project:
> - Install OpenStack
> - Deploying OpenStack Sahara source
> - Deploying Mahout on Sahara OpenStack system.
> - Construction of the API.
>
> Through web or mobile interface, users can:
> - Enable / Disable Mahout on Hadoop cluster
> - Run Mahout job
> - Get information on surveillance systems related to Mahout job.
> - Statistics and service costs over time and total resource use.
>
> Definitely!!! APIs will be public. Look forward to your comments.
> Hopefully in this summer, we can do something together.
>
> Thank you very much! :)

dat,

since mahout is a great ml library that leverages mapreduce (and now 
spark & h2o), it may be simpler for you to make sure that mahout is 
installed by the various sahara plugins.

in fact, i bet you could run mahout jobs using edp and the java action 
right now in sahara. if that's true it's probably a bit clunky and worth 
the effort to streamline.

best,


matt



More information about the OpenStack-dev mailing list