What are IOPS (Input/Output Operations Per Second)? How are they measured?

How is IOPS measured, and what does it really mean?

IOPS is heavily used as a benchmark for storage devices and networks. It helps describe various types of storage devices, including solid-state drives and hard disk drives. Additionally, it can gauge the speed of a Storage Area Network (SAN).

So, what do IOPS mean? How do they work, and how are they measured to indicate the speed of a storage device? To understand its meaning, we need to delve into and analyze how it is measured. Let's get started.

What do IOPS stand for?

IOPS stands for Input/Output Operations Per Second. It measures and helps understand the read and write speed of a storage device. In tech circles, it's often pronounced as "eye-ops" and indicates the speed of computer storage operations.

Memory types commonly checked by this unit include:

• SSD – Solid State Drive
• HDD – Hard Disk Drive
• SAN – Storage Area Network

In other words, it's a benchmark unit that ensures your device's storage speed matches what the manufacturer claims. There will be some variance in these IOPS figures since manufacturers test those speeds in ideal conditions.

And so these benchmarks very rarely translate to real-world usage, like applications, gaming, file copies, and so on. And so when you actually look at IOPS in a live situation, it's always a fraction of what the manufacturer is actually advertising. Here's an example.

As you can see, the read and write speeds are 6684 and 4918, respectively. However, the manufacturer claims that this SSD should write at around 7000 MBPS and read at 5000 MBPS. So, where does this difference in IOPS come from?

The answer lies in the manufacturer's claims, which specify the maximum write speed at 7,000 MBps and reads at 5,000 MBps. It might achieve those speeds under ideal conditions, but in the real world, what matters more for performance is IOPS (input/output operations per second).

IOPS Performance Characteristics

IOPS (Input/Output Operations Per Second) is a straightforward metric, and its meaning can be easily understood. To provide more insight, here are some common types of IOPS measurements in storage devices: 1. Sequential Read IOPS: Measures the device's performance when reading data in a continuous stream. 2. Sequential Write IOPS: Measures the device's performance when writing data in a continuous stream. 3. Random Read IOPS: Evaluates the device's capability to handle small, unordered data requests. 4. Random Write IOPS: Evaluates the device's capability to handle small, unordered data writes. 5. 4K Read IOPS: IOPS for 4KB block size reads, often used to assess system performance with small files or database queries. 6. 4K Write IOPS: IOPS for 4KB block size writes, also used for evaluating small file or database write performance. 7. Mixed Workload IOPS: Reflects the device's performance under a combination of read and write operations. 8. Burst IOPS: Represents the maximum operation rate a device can achieve over a short period. 9. Sustained IOPS: Measures the stable operation rate a device can maintain over an extended period under high load. Understanding these measurement types helps evaluate and compare the performance of different storage devices.

If you characterize these storage devices with a common tool, a disk benchmarking tool on your PC, it would look something like this.

As such, to measure IOPS, you need to know the characteristics of your computer's hard drive. That way, you can interpret the results given by these IOPS tools in terms of the overall speed of the physical drive.

How do you measure IOPS?

You can use the Iometer tool to measure IOPS easily. It's a powerful I/O subsystem measurement tool that lets you characterize storage drives in both single and clustered computers.

Advise users that they can use a tool such as Iometer to measure IOPS. However, there can be significant differences in IOPS performance for SSDs, HDDs, and SANs. For example:

- HDDs are rated using standard computing metrics, such as IOPS (input/output operations per second). - SSDs perform based on their internal controllers, like DRAM caching or the NAND controller. - HDDs have higher latency and slower read/write speeds. - SSDs are much faster with sequential read/write operations, resulting in better performance for real-time applications.

So, how do you measure the actual IOPS of storage? You'll need to download a disk benchmarking tool such as Iometer.ORG or CrystalDiskMark. After downloading, you would then:

Step 1: Download and install the program

Second step: Run it.

(Picture: Crystal Disk Mark - Results of all tests)

Step 3: Select the disk for which you want to measure IOPS.

Step 4: Let the tool do the measuring

Step 5: Analyze the final reading result

As you can see from top to bottom, there are various readings of input and output per second, as we said before.

• All = Total IOPS
• SEQQ1M and SEQ12K = Sequential reads and writes
• RND4k = Random reads and writes

So this tool allows you to do the inspection in the MBS, but there's a little trick: you just have to

Step 1: Click on the menu in the top right corner.

Step 2: Choose IOPS over MB/s

Now it will show you the IOPS measurement:

Alright, next on the list is the Crystal Disk Mark read and write test results: 

So it's very easy, with any Iometer tool, to find out what are the IOPs and the MBs or GBs per second of your storage device.

IOPS and Throughput

Bandwidth is a measurement of a storage device that tells you how much data a hard drive can transfer. Compared to IOPS, bandwidth is more relevant for streaming services, such as image hosting sites and video files.

Because it is a continuous measure of storage in transit, it is well suited to streaming media and video.

On the other hand, IOPS is the primary measure of storage device speed, as seen in this article. So, what's the difference between them?

While both measures are common, IOPS provides a better sense of the speed of a physical storage device. Throughput, which is more relevant to industrial applications, is used by large enterprises, streaming services, and data centers.

Bottom line

We hope this helps you understand the importance of IOPS and how to measure it. Not only do we delve into the significance of IOPS, but we also introduce its counterpart, throughput. So, leverage this article to gain more insights into IOPS and measure it with the tools we recommend.