Hi Joe,

The planning of hardware, software and cluster planning depends on many factors like:
- what kind of data you'll be prosesing with this cluster
- rate of data that is being ingested , also called as Velocity
- how many mappers and reducers should be executed in each nodes and many other factors.

The decision on what kind of hardware , software and size of cluster
Will be decided on some of the above factors and there are multiple stakeholders to decide along with also keeping the financials in mind.

Thanks