Comment by kragen
6 hours ago
Latency and throughput are not interchangeable, and it is entirely normal to do millions of subroutine calls per HTTP request, so even if latency is your only concern, a subroutine call operation with a latency ten thousand dollars faster than the HTTP request might still be too slow.
Latency is not interchangeable with throughput because, if your single 8-core server needs to serve 200 HTTP requests per second, you need to spend less than 40 core-milliseconds per request on average, no matter whether the HTTP clients are 1ms away or 1000ms away.
"ten thousand dollars faster"?
Yes, at the abstraction level of money, many more things become interchangeable, but in order to have something worth spending money on, we have to distinguish between those things.