And why you need that so fast? Consider that there are several reports and all are FIFO, no parallel transfers that is why there are some delays with some reports.
@cav - I think I have and idea why “so fast”…
@janegriff: Your example is a slightly flawed, since it only shows 1 chunk of data passing through the ETL. This is what is actually happening:
00:00 begin time (chunk 1)
00:15 update time (chunk 1), begin time (chunk 2)
00:30 receive time (chunk 1), update time (chunk 2), begin time (chunk 3)
00:35 transfer time (chunk 1)
00:45 receive time (chunk 2), update time (chunk 3), begin time (chunk 4)
00:50 transfer time (chunk 2)
…etc…
…you get the idea.. it’s pretty busy around the 15 min aggregate… and that’s without the hourly, daily, weekly, etc…
I think my point is that this is historical reporting and your apparent urgency to display the results implies that you are attempting to use the Solution for near-real-time reporting. I really would recommend that, if you need real-time results, you use CCPulse and perhaps plug it in to your Historical reporting, to fill the gaps. i don’t believe you will get the results any faster, unless you reduce the number of Objects of Statistics being gathered…