lowkey saw a cool breakdown of smp like postgresql or oracle versus the big ol' distributed systems in
mpp land. on one side you got those single-server champs perfect for transactional speed, then across to these multi-node wonders designed specifically for massive analytical queries. i mean if u need super fast transactions and microservices stuff - go with an smp db like postgres or oracle; but flip the switch when your data volume hits 32% increase point (no exaggeration) in analytics workloads where youre dealing terabytes of info.
anyone else had a wild ride moving from one to another? im curious about how handle that transition!
more here:
https://hackernoon.com/traditional-vs-mpp-databases-architecture-scaling-and-workload-tradeoffs?source=rss