As of version 6745 it is possible to execute multiple Deliver instances at the same time on the same server. This feature is great as you don’t want to process 30+ semantic models one by one.
There are still enhancements to be made the parallelism functionality:
- It should also be possible to process multiple DWH/Prepare instances at the same time. We have about 8 production Prepare instances that sometimes have to wait for each other to be ready.
- For using parallelism, all SSL execution packages should be in the same Job. When you’re using Exmon for process orchestration, it is nice to see the status of each individual semantic model on you process map. When you put all SSLs into one Job, this is not possible anymore. You will see a red light if just 1 out of 30 models failed to execute, which gives business users wrong/too few detailed information (and causes unnecessary panic).
- Have control over the order of execution. Some SSLs need to process first, as the business are waiting for them to be ready each morning. Others can run later. Hence it is important to set priorities and have control over which SSLs will execute first.
