Model high-write, horizontally distributed workloads with wide-column databases. Design partition and clustering keys around known query patterns.
