Both IOPS and MBPS are standard benchmarking measurements when testing the speed of a storage device.

IOPS-I/O Per Second is nothing more than the number of I/O operations I can perform in one second. For applications you might be able to get information about the number of IOPS requested. An Oracle database has the following internal statistics for IOPS:
Total Reads :'physical read total IO requests'
Total Writes:'physical write total IO requests'

MBPS-Mega Bytes Per Second is classified as throughput of a storage device and equates to the Bytes Per Second a storage device can stream. Likewise an Oracle database has the following internal statistics for MBPS:
Total Bytes Read :'physical read total bytes'
Total Bytes Written :'physical write total bytes'

Both IOPS and MBPS definitions given above are in their 'purest' form. I say purest form because many things can influence the number of IOPS or MBPS you might see on a storage device. While we would like to all have fair testing practices, things such as testing only the outer portion of a disk, artificially increasing queue depths, or adding cache can make IOPS numbers skyrocket and make a storage device look better than it is.

For fun, this week (it might take more), I will, among a few other posts, walk you through looking at IOPS and MBPS in an Oracle database. Certain aspects will be very familiar to storage administrators and some very familiar to DBAs (if they are reading this blog). But what I hope to give is a common ground both groups can come together to validate the performance of their storage/database system.