Launch mahout using hadoop on Amazon EMR / EC2

Question

I want to transfer my current local cluster to Amazon. In this cluster, I use services such as mahout, hbase and hive. I have two options now on Amazon, either go to clean EC2 instances, or reduce the cluster on the Elastic map. I want some suggestions on what is the best option for moving a cluster that has such requirements.

Answer 1

I always suggest people go for EMR, because it is managed and will be a bit more expensive than using pure ec2, but the headache and time you spend setting up clusters and then managing them can be saved by running a managed one such as EMR.

• Mahout can be easily run as a custom jar .
• Hive Cluster can also be launched in minutes.
• Similary for HBase, Amazon recently added the creation of an HBase cluster through EMR.

See other views here .