Is the BPM necessary for the RosettaNet implementation on webMethods? The BPM performance rely on the System DB performance, so it will be a big challenge for the DB performance for volume transaction. So would you please share your experience on this? Thanks in advance!
my understanding is that you need to use Processing Engine to execute your process model for RN. Processing Engine will need the DB tables to track the status of the model.
So, even you don’t want to view the model status in MyWebmethods (BPM), the DB tables are still neccessary.
You can change the level of the logging/pipeline savings to avoid unnessary DB activities, and also do some performance tunning on the DB side (adding index for example)
BPM performance rely on your implementation.
DB can become a bottleneck depending your workload and your audit/logging settings.
But as you can scale the number of IS’s in a clustered process runtime, and a MWS cluster, you can also scale databases.
At the end you need to set a proper QA environment, identify bottlenecks and take proper actions/solutions before moving to production.
Thanks all for your reply.
The RN data will bu submit to TN firstly, and then trigger the process models. So I think it is possible to create a TN processing rule and invoke a customized service. That is, just process the RN data as a XML. So is there some has the successful experience on this?