Expand my Community achievements bar.

SOLVED

How to improve the concurrent compute capability of workflow?

Avatar

Level 3

Hi Team, 

Could give us a guild line or a consult about this topic?

For we encountered a situation, that we are provided a interface for third-part data import.

For the reason of third-part development ability, they demanded us to provided a workflow to import its data. But it seems a high-frequency execution workflow, which has two activity node like below. The two node is js for data process and xtk.session.Write(myXML). But we found that, if the external signal invoked in 50 time/second, the workflow would be very slowly. Is there any method to improve the performance of workflow?

yueminspb_0-1610692373659.png 

yueminspb_1-1610692455834.png

yueminspb_2-1610692518255.png

 

 

1 Accepted Solution

Avatar

Correct answer by
Community Advisor

Hi,

 

ACC web services call have some limitation on TPS. It's not recommended to do bulk API update.

Work around can be:

  • the API write data to a file and a workflow pick up this file base on a frequency.
  • the API write data to a database and ACC connect to the DB to process the data
  • the API send requests to a middleware or microservices and this middleware will manage the queue /traffic to ACC

Thanks,

David



David Kangni

View solution in original post

3 Replies

Avatar

Correct answer by
Community Advisor

Hi,

 

ACC web services call have some limitation on TPS. It's not recommended to do bulk API update.

Work around can be:

  • the API write data to a file and a workflow pick up this file base on a frequency.
  • the API write data to a database and ACC connect to the DB to process the data
  • the API send requests to a middleware or microservices and this middleware will manage the queue /traffic to ACC

Thanks,

David



David Kangni

Avatar

Level 2

You may consider an approach based not on external Signal / SOAP based workflow, but a JSSP with the connection handling and db write in one script.

 

The JSSP executions will multi-thread and scale to the number of available db connections.  You can then scale frontal servers/web modules as required.