In the Essentials it appears that MapR has been built to improve upon the Hadoop Open Source performance, reliability, and support POSIX (e.g. standard unix/Linux commands etc). One thing it states is that MapR components are "fully compatible with the Hadoop and its ecosystem components". So, I was just wondering a few things:
Question 1: What effort is involved in migrating to fully to a MapR solution from a general open source Hadoop solution.
Question 2: What effort is involved to migrate back to a fully open source Hadoop solution from MapR.
Question 3: Can pieces of the MapR components vs full MapR components exists in a hybrid like environment?
Question 4: Would we expect eventually that the Hadoop Open Source project will implement MapR features/functionality?
My thinking around these questions is about the costs involved (capital and ongoing support) for moving to a commercial product vs purely open source. There may be a time in my application lifecycle of requiring more robust support with a vendor product, and then towards the end of the lifecycle reducing support costs.