Requesting basic architecture advice, please

Question

Hello all,
I've got a situation where a customer will make a request of a web app and needs to receive information back from the web app anywhere from 30 minutes to an hour later. This is well beyond the length of a reasonable HTTP timeout, and even if I extended the timeout and kept the session alive, I'd be worried about running out of resources as this will get many, many hits.
Can you suggest an approach that might work here, please? Would you put an app server at each end and just pass http calls to servlets back and forth? An ESB with two clients so they can pass messages back and forth?
Thanks in advance for your time!
Rgds,
Bret

Answer 1

You will want to have the initial request be asynchronous. Simply return a response to the client that the request was properly formatted and is being processed. From that point, you have a few options:
1. Use email or pager notification to let the user know that the request has completed
2. Use AJAX or something similar to update an area of the page when the request has completed
Regardless, you will be passing back to the user a token of some kind. When they request the results of the report/etc. they will furnish the token, which you can use to fetch the results that were previously calculated. You might also have a need to clean out the results of processing, say overnight, if the results should not persist indefinitely.
- Saish