Cassandra cluster and several different accessing systems

[Jan Algermissen <ja...@nordsc.com>]

Hi,

I am designing a larger system that is composed of several independent smaller systems (think "services"). Each of the smaller systems has its own funtionality, logical database, and data access patterns.

My plan is to use a Cassandra cluster as the ("cloud") database solution for all/most of these systems and I wonder how I should set this up.

The individual system owners should not need to worry about the database beyond setting up the key spaces they need.

Should I design a single cluster and let the system create and configure their key spaces as they see fit? Or should I design 2-3 independent clusters tailored towards certain kinds of access patterns and replication/consistency needs (focus on read , focus on write, focus on amount of data or simliar) and then constrain the kinds of key spaces allowed for each cluster kind?

Or should every system be given its own cluster?

Or did I maybe miss another option?

Jan