“divide-and-multiplex”

https://engineering.mongodb.com/post/multiplexing-golang-channels-to-maximize-throughput/

How to marry high throughput concurrent workload with a serial output. Mike Thorsley points out that it's similar to Spark's group by.