Either: 1) Latency/peformance/concurrency/server costs or 2) Extreme correctness If neither of these things is a current or anticipated problem for you, it

Either: 1) Latency/peformance/concurrency/server costs or 2) Extreme correctness If neither of these things is a current or anticipated problem for you, it may not be worth it (though you could always do an exploratory project to find out!). Make no mistake: like any engineering decision, for all its advantages it does come with some costs.