All of these solutions are still battling the fact that distributed systems are hard and there's a tradeoff to every decision. If you use synchronous replicati

All of these solutions are still battling the fact that distributed systems are hard and there's a tradeoff to every decision. If you use synchronous replication, you now require at least two systems to commit a write, so you've increased the chance that your commit will fail. Maybe that's the right choice—but it's a tradeoff and the decision ultimately rests in your business requirements. My point is simply that there is no magic bullet in ensuring that all relevant systems have the same answer to the same question at the same time.