I need to bring an API for an old existing product to the web. It will be a REST API written in C#. I need a process via API call because my existing library is definitely single-threaded. But of course I could reuse a process. So I thought of a kind of process pool in the background, which processes the API calls.
Basically, I have two ideas: Is there a way to configure Kestrel as a web server for this? If it would use a process for each HTTP request, it wouldn't be perfect yet (the initialization of the API is slow and 32-bit only) but a good start.
The only alternative I could think of would be to manage a pool of worker processes myself and transport the tasks and results there via memory-mapped files....
None of this is really nice, but I have to deal with the old API somehow. Do you have any other ideas or how would you do it?
Thanks for your help.
Ich habe schon ein bisschen mit den Kestrel-Optionen rumgespielt, wobei da nichts richtig Sinn macht ausser: AllowSynchronousIO = true;