Wanted to know can PI Server database is efficient for storage and retrieval of archive data over unlimited period of time. Assuming hardware to be the only limitation.
Thanks in Advance,
Thanks & Regards,
PI Server is made for storing and retriving data for long times. I have an PI Serve rrunning that holds data from 2001 till now. I have about 700GB of archives now. There is no real limitation - current Server Hardware is performant enough (asuming you have an Server CPU and a RAID System). If you have realy heavy access to your data you can optimise mostly if you change how much archives are cached in RAM, so access is faster. There are other things which can be optimised (network speed; disk I/O, CPU threads) but in my experiance on current systems this should be no problem.
You can post some figures how much data (TAGS, scanrate, compressingrate etc.) you asume for your system.
Great response, thanks fLOSt.
The PI Data Archive software has very high tolerances for data storage and retrieval, with sites up to millions of PI tags. As mentioned above, the real limitation would be hardware, which will depend on your specific system needs: Tag count, scan frequencies, compression rates (archiving rate), and the various performance metrics that you deem appropriate.
If speed is a concern, we will highly recommend hosting your archive files on SSDs so that you're not held back by disk speed limitations of spinning disks, and the amount of RAM on the system is important for caching data for instant retrieval.
If you need to store historical data before 1970 you should test. In the past there have been some capabilities issues with data displayed in DataLink and ProcessBook. Anyway, I don't think that is your question, just wanted to note a possible limitation.
I think you should check these videos first...
OSIsoft: What are Archive Files & How Are They Used to Store PI System Data? - YouTube
to understand how PI differs from traditional SQL database.
I was just going to post this link. It doesn't go too deep, but it explains pretty well why the data retrieval doesn't get slow over time.
Retrieving data ...