A inter-instance ping is when one instance requests a data block
which has been modified by another instance but not yet written to disk.

OPS writes or "pings" such block to disk so that the requesting instance can read the most current data.

If all OPS-instances modify the same data blocks, "pinging" becomes very I/O-consuming.
That is why data and application partitioning between instances are so important.

(In OPS 8i, the new feature Cache Fusion directly transfers data blocks among instances
eliminating excessive I/O, but data partitioning is still quite important for OPS performance.)