In my understanding, if it takes a long time to get response from server (for example high-latency calculation is needed), we'll go with asynchronous way that 200 response is first generated while the underlying calculation is still going on. We could even put the request into kafka queue to achieve the goal.
But we also have long polling that we simply keep stable connection until we get response from server, which allows sufficient time for high latency.
Then what's the difference between these two approaches? Also please correct me if any part of my understanding above is wrong. Thanks