We have a need to reduce time between two screens which are chained. For t

rajeevj over 9 years ago

We have a need to reduce time between two screens which are chained. For that we tried to do a parallel processing for things which are not dependent to user input. Attached snapshot is a child process and submission of this process will invoke another process which is again having a user input task. Thus there are 3 processes having user input tasks forms in sequence and attached snapshot is the 2nd process in sequence.

As can be seen in the attachment there are 3 parallel path from Start Event i.e. i) Fetch HO_App specific Data ii) Set Progress Indicator Array iii) Get Data. These activities are consolidated at “Complex” node, as nodes after that are dependent on the completion of the execution at previous nodes.

Problem that we are facing with parallel execution approach is that the “Chaining” is breaking. We have tried all gates in place of complex gate, as well as we have tried to wait at that point using a “Rule Event” but with no ...

Discussion posts and replies are publicly visible

0 rajeevj over 9 years ago

...success.
Please suggest how do we achieve the parallel execution to utilize the time that user is going to spend on task form.
Cancel
Vote Up 0 Vote Down

Sign in to reply

Verify Answer

Cancel
0 sumitp584 over 9 years ago

if the two tasks are not dependent on each other, why dont you put it in two different sub process models and run in asynchronous..
If you want the node to wait till everything comes in, use a blank script task rather than complex gate.
Am just assuming your problem here, if am wrong could you please elaborate what you are trying to achieve here.
Cancel
Vote Up 0 Vote Down

Sign in to reply

Verify Answer

Cancel
0 PhilB
A Score Level 1
over 9 years ago

Have you looked into resolving the performance issue rather than working around it by using parallel processing? What are the nodes doing? Simply setting some variables and retrieving some data should perform just fine when chained; maybe there's a more serious issue with the data or query rules which should be fixed rather than avoided?
Cancel
Vote Up 0 Vote Down

Sign in to reply

Verify Answer

Cancel
0 Jason Ruvinsky
Certified Lead Developer
over 9 years ago

If you aren't worried about a race condition, you could put the "parallel" nodes into an asynchronous sub-process on their own. The chain is broken because you're waiting for unchained nodes to complete at the Complex gateway.

You could also look into the expressions on those nodes that are causing the performance issues and see if there is a way to improve your performance there.
Cancel
Vote Up 0 Vote Down

Sign in to reply

Verify Answer

Cancel