I’ve got an application written in groovy. It takes some cmd args and returns previously formatted response. As system grew, it appeared that it is required to run this app extremely frequently (like 80 times in 5 mins) which leads to certain performance issues. In particular it creates all its objects over and over again which leads to filling up to 60MB RAM in one run (can be easily calculated how severely ROM/swap is used).
I want to migrate it to a service running mode which will simply take certain params and return formatted output. But:
- App is always triggered by a bat/sh script (this can’t be changed)
- Both script and app are on the same host server
So, I’m wondering how it would be better to perform the communication of a script and a service?
P.S.: Sorry that I didn’t mention, it’s a standalone app, it will never use a server or anything like that as it appears to be redundant. Solution should be as simple as possible and extremely lightweight.
Example: The simplest thing I can think of by now is never to migrate it (I know it’s contradictory ;)) and simply introduce a DB where all thee result will be stored and an app will have it’s own schedule of when to trigger. Whenever it is triggered with any params, it should simply search the latest result in DB and return it. Easy, light, fast, and working. 🙂
For enterprise environments I would suggest a JavaEE application with EJB running in an application server. For your requirements this might be an overkill. A simple solution can be:
RMI Server
RMI Client
RMI Interface