RE: PI Archive 2015 R2 with AF 2.8. 1000 elements, 6000 attributes, 50 events/day, 5000 tags
We've been building out our AF model over the past year mapping it to our PI Points. Generally speaking, we've kept it simple and done a simple "1:1 mapping" data reference from an attribute to a PI Point. We like the benefits AF offers, but came across a comment in the OLEDB Enterprise manual that is prompting this inquiry. Traditionally, we've used OLEDB to make SQL calls to extract PI Archive data. We were planning to "upgrade" to OLEDB Enterprise to do roughly the same thing, but instead hitting AF.
The OLEDB Ent. manual indicates that accessing archive data is faster if you use OLEDB. However, the OLEDB manual also indicates that 2016 is the last version they plan to release.
I'd really rather use just one connection (OLEDB Ent.) to include data and events. But, if performance isn't ideal this would make me rethink our approach.
Does anyone have experience extracting archive data using OLEDB versus OLEDB Enterprise? Is performance comparable?