[privacyfornow]

One way to understand CQRS is to realize that even the traditional model relies on CQRS underneath, it just so happens that the relational databases hide some of this behind a blocking wait on commit while a background thread reads and flushes the event log (consisting of SQL statements written to a redo log for example). Really what you are gaining by inverting the database like this is a lot more flexibility and simplicity for initial development and subsequent change management (maintenance). The flexibility I speak of manifests in your ability to tweak availability by introducing controllable circuit breakers with availability at the expense of some latency or alternatively putting an upper bound on latency at the expense of some availability (with every tradeoff fully documented and demonstrably adhered to).