As far as I know, in Akka, all participants are planned by the thread pool. Too many participants block IO at the same time, each blocking call blocks one thread, which causes the thread to shut down.
Now I'm looking at an interesting fiber implementation on the JVM - Quasar , which allows many custom threads - fibers - and uses a thread pool to plan them. However, I am wondering if there are advantages when many fibers call legacy blocking APIs.
I donβt think this would help, because the blocking API would still block the system thread, and Quasar could not magically turn it into just a fiber lock. This is just my guess, please correct me if I am wrong.
source
share