AnsweredAssumed Answered

Alternative Uses of OLEDB Enterprise

Question asked by skermode on Jun 5, 2013
Latest reply on Jun 6, 2013 by drabinowitz

First off: hello, world!


Okay, now to business...I'm working out a plan for a PI system where the end users like the idea of using PowerPivot to access cubed PI data from the OLEDB Enterprise platform, but want that exact same cube data available to other systems.


I have some background with the PI historian and ProcessBook, but not nearly as familiar with this platform, hence the questions.


Imagine the following scenario:  end user is attempting to target some specific business intelligence over a recent time period and would like to slice based on the entire company's data, a specific region, and then a specific site.  Based on the seminars I've watched, OLEDB Enterprise connected to Excel PowerPivot will work very nicely.


Now, take it a step further:  end user wishes to use some custom, in-house software that will query the cubes much like PowerPivot would, only for other applications (ex. web).  CoreSight is an option for some of the data, but the end-user is very excited about having access to the cubes directly without having to go through the historian.


And now the questions...


  1. With OLEDB Enterprise, can a "cached" SQL environment of archived PI data be configured?  Of course *anything* is possible with enough engineering, but I'm more so targeting an out-of-the-box type solution versus custom.
  2. Assuming the SQL cache-cube environment is possible, what type of "lag" could we expect from the more recent archived data point in the cube versus the current value stored in archive?  Realistically, what's the probability of getting the "now" value in the cache every minute in a 1,000 tag environment? 10,000?  100,000?  1,000,000? Gut feelings are definitely acceptable here.
  3. Again, assuming #1 is something possible without designing the cube architecture from scratch, how would business users design new cubes without a DB admin helping them?  Obviously, we can trust only a few, power users to perform these tasks--but we don't necessarily need to send them to SQL school to perform them.  Again, if #1 has to be en entirely custom system, then the answer to this is simply, "go write a custom cube-maker."

Disclaimer:  I currently do not have access to an OLEDB Enterprise system to sandbox with.  Once I can leverage a PI-based solution, I can obtain the license and start tinkering.


Thank you!  I value any input and I will do my best to help explain any details that you feel are necessary.