• assigned: fredm
• type: feature
• priority: high
• keywords: correlations, asynchronous jobs, external process
• status: completed, closed

Currently on GeneNetwork2, we run the computations in an external process, but the application keeps the request alive until the entire computation is complete, and the results parsed.

For a lot of the computations, this is not a problem, but there are cases where the computation and data processing takes a huge amount of time, leading to timeouts either on the back-end, or on the browser end. This is a

• bug

I (fredm) propose to implement a "Request Architecture" that works as follows:

• STEP 01: User requests a correlation computation
• STEP 02: The application triggers an asynchronous external job and responds to the user with the job identifier
• STEP 03: After a short delay, nhe user's browser uses the identifier to query the job status
• STEP 04: If the job is not completed, go to "STEP 03" above, otherwise go to "STEP 05" below
• STEP 05: If the job completed with no error, go to "STEP 06" below, otherwise jump to "STEP 07" below
• STEP 06: Process the results and produce the HTML page to display to the user. Jump to "STEP 08"
• STEP 07: Process the error and display the errors page. Jump to "STEP 08"
• STEP 08: Correlations computation process is complete

----

This has been completed.