I am looking at JBPM as a candidate workflow engine. My current understanding is that if I fork a path of execution into multiple concurrent paths, all of them will be executed in one thread (for each path: until it reaches the join node or a wait state) - sequentially.
I need to understand what are my options if I want to execute them in parallel - i.e. in multiple threads (I can bring my own pool for the purpose). My usecase is: I need to execute lots of (independent) queries against (multiple) databases with varying (sometimes bad) response time, while keeping the runtime very low and the workflow definition readable by a Business Analyst.
As far as I understood async forks will be executed in different threads. For each activity after the fork a job will be created by the JobExecutor which maintains a pool of threads to execute jobs. However I am not sure if I understood your question correctly. If not please rephrase and clarifiy.