here I will expose a need I think that is really common when explaining PI/AF to customers: PI/AF performances.
From time to time we get some hints of information, but I would really like to know what are the maximum tested performances of PI and AF.
I know that it will depend on test machine and on source data system, but I suppose OSIsoft performs stress tests on the PI system and will have some information about the following items:
- Max PI Point count: I recall a 20Mtags with PI Server 2012. Has it changed?
- Max event reads from PI: last information I have found is 10Mreads/s
- Max event wiites to PI: last information I have found is 1Mwrites/s to snapshot and 500kwrites/s to archive
- Max buffered events throughput: last information I have is 50events/s
- Interface fastest scan class frequency: I know it depends on the interface, but for what refers to PI OPC DA Interface (for being the most common)... can we get to 1ms scan class?
- AF Analysis service: calculation frequency: with a monster server, what is the max expected throughput (supposing simple calculations)
Thanks in advance