"Best" ExecutionContext for IO

Question

I have some synchronous calls in my Scala code. I've wrapped them in a blocking() context and then in a Future: Future(blocking(syncCall())), but I don't know which type of ExecutionContext to use.

I know there may be a lot of possibilities and that there is not an ExecutionContext that it's the "best". I just need some information in order not to choose the worst because there is a lot of information out there and I have a mess in my head.

Pontonier

Why are you wrapping the call in a blocking context if you're running this inside a Future. Why not Future(syncCall())? – Boson

@YuvalItzchakov to notify the ExecutionContext that it's a blocking call, read github.com/alexandru/scala-best-practices/blob/master/sections/… – Pontonier

Good read, thanks. Following that section (4.6) says: Related to Rule 4.4, if you're doing a lot of blocking I/O (e.g. a lot of calls to JDBC), it's better to create a second thread-pool / execution context and execute all blocking calls on that, leaving the application's thread-pool to deal with CPU-bound stuff.. Why not use that advice? – Boson

@YuvalItzchakov that is my intention and this is why I'm asking for some help on that, because I don't know which thread pool to use. Also, 4.6 does not invalidate rule 4.4 – Pontonier

Scala's ExecutionContext doesn't do any thread scheduling on its own, but rather passes the job to the underlying java.util.concurrent.ExecutorService. So your question boils down to the following: which type of ExecutorService should one use for IO tasks. – Craigcraighead

I would suggest go with section 4.11 in the reading above. Introduce few system properties where you can define your thread-pool: maxPool size, initialPool size etc. – Jamie