Processing parallel multiple database lookups using 4GL??

I work at a site that has a distributed network of 9 distributed Informix databases. Each database has information that is unique to that database, plus global lookup and validation tables. Each database has the same schema. The problem: We have jobs that run across all databases. Each database can take from 5 minutes to 20 minutes to process. Because of size restrictions on "Select" statements, we have chosen to execute individual "Select...INTO TEMP" statements to extract data to temp tables, and then combine the temp tables into the final report or extract. What I would like to do is execute all the Select statements at one time, and then coordinate it so that the parent process waits until all of them are done before it does the final combination. While this should consume roughly the same amount of CPU time as the original, the results will be available much quicker. I have some rough ideas on how to do this, using control tables and the 'RUN' command to execute multiple child processes. But before I expend a lot of time and energy re-inventing the wheel-- 1. Has someone successfully/unsuccessfully attempted this, and what pitfalls await me. 2. Would someone be willing to part with an algorithm?? If no one answers the above requests, I would like to propose developing such an cooperativly here, live, and sharing the results. I would be willing to chair such an activity and coordinate the efforts. I humbly await for insight from the ether. John Callaway Bath Iron Works The opinions stated above are my own. If you chose to place any value on the above opinions, I strongly suggest you first pipe them through sanity.check.