Hi John!
I asked a similar question in the dev mailing list some days ago, so if you want to take a look here's the link:
Stefan and Rainer helped a lot in clarifying some doubts (thanks again!). The goal would be to eventually add more examples and references in the official documentation, so any input from you would be really valuable. Feel free to suggest what you would like to see in the docs :)
My view of the AsyncRequestWorkerFactor is how much overcommit you are willing to accept for your workers to maximise throughput without risking too much (too many connections in the listen queue, timeouts, keep alives dropped, etc..). Raising the Async factor to an higher value might bring throughput increase but there is always the risk of overcommitting too much; the final value highly depends on your traffic pattern, so careful testing is the only good way to improve performances. Rainer explained this concept in a much better and complete way, I would suggest to check it first.
Let me know!
Luca
