I understood the concept of buffering as it was well explained in the course. However a question that just come up to my mind is: How the Buffering mechanism works in case of interface with fail over or high availability?
In short: With n-way buffering, the buffering service queues data independently to each server in a PI Data Archive collective.
As Roger indicated, when a particular instance of PI Buffer Subsystem is sending data to more than one PI Data Archive, it has a separate queue for each recipient Data Archive. The most common time when this will happen is when you're using the PI Buffer Subsystem to send data to all members of a PI Data Archive Collective (we call this n-way buffering). However, it's also possible that you have interfaces sending data to entirely unrelated PI Data Archives on one node (and thereby one instance of PI Buffer Subsystem).
There's really only a small amount of interaction between the PI Buffer Subsystem and interface failover. If you're interested in getting at how to maintain data fidelity during an interface failover event, I'd suggest we have that conversation separately. For a slightly longer answer here: the PI Buffer Subsystem only really comes into play once an interface is sending data. Whether or not interfaces are attempting to send data, and under what circumstances they attempt to send data, is set by the failover configuration.
Retrieving data ...