I can't really understand the idea of installing "monster" software onto tiny machines. I wouldn't be totally puzzled if this post came from some hobbyist, but it is released by Basho itself. So what's the idea?
These "tiny machines" are still massively larger/faster/more flexible than the enterprise workhorses of not that many years ago...
The challenge of getting "monster" software onto them is an interesting exercise in part because it reveals a lot about the software when you see how it acts on machines that are constrained and slow by modern standards.
I feel I still learn far more from working on platforms with severe resource constraints than elsewhere, because a lot of issues that still affect larger systems but that are masked suddenly become massive in-your-face challenges.
You could take advantage of the replication for some kind of highly reliable monitoring device: take different samples with each raspberry pi, the data is duplicated on N devices, and then you can come back later to collect not worrying about sd card or device failure.
"I thought it might be fun to create a Riak cluster on a set of Pis I had lying around... Note that cluster isn’t fit for any purpose in its current form."
The challenge of getting "monster" software onto them is an interesting exercise in part because it reveals a lot about the software when you see how it acts on machines that are constrained and slow by modern standards.
I feel I still learn far more from working on platforms with severe resource constraints than elsewhere, because a lot of issues that still affect larger systems but that are masked suddenly become massive in-your-face challenges.
And of course it can simply be just fun.