Input/ Output per second (IOPS) is a standard measures read and write performance of disk transactions. These are the read and write operations per second recorded at the block layer between the host and its storage device(s).
IOPS provide a common performance measurement used to benchmark rotational (hard) disk drives (HDDs), Solid State Drives (SSDs) and storage area networks (SANs).
Optical Prime measures IOPS in the form of read and write I/O and displays these metrics as a stacked graph over a period of time. Transactions are measured simply as the average count of read I/O plus the average count of write I/O per second. Read I/O is an operation that is conducted when the host operating system requests some amount of data from disk and the storage device retrieves it. Write I/O is an operation where the host Operating System commits—or saves—data to a storage device.
Because Optical Prime is a host-based performance measuring tool, array-based RAID is not relevant. Host-based I/O is often called front-end I/O and is, thus, unconcerned with storage caching, RAID layouts and striping.
Recording host based-based IO creates a universally understood measure of requirements by providers of various modern technologies. Each of those technology providers might handle I/O uniquely inside their product or platform, however, host-based I/O is what your server believes its performance demand to be.
As a general rule of thumb IO Size, the number of IOs, and Throughput have a dependent relationship
Avg. IOPS = Throughput / IO Transfer Size
Or
Throughput = IOPS * IO Transfer Size
Related To:
Throughput
Block (IO) Size
Queue Depth