Three Common Key Storage Performance

There are three common key storage performance metrics for all environments: Latency, throughput (or bandwidth) and IOPS.

Latency (also known as response time) describes the time taken to complete an I / O (input output) single operation and is basically a measure of how fast the storage system responds to reading and writing requests.

Values ​​are usually measured in milliseconds, with the fastest drive now citing a fraction of a millisecond. In the ideal world, the latency will be zero, without penalty for the application to read / write operations to the permanent storage media. However, the physics gets the path and there is latency for every I / O operation.

The purpose of storage solutions is to minimize the value of latency, since storage is often an impediment to IT infrastructure. Low latency means less waiting for I / O refinement and therefore results in faster execution.

Throughput – The storage system’s ability to transfer the amount of fixed data within the measured time is known as throughput, or broadband. Typically, the throughput is measured in megabytes per second (MBps) or similar units.

Storage instructions and disk devices can be measured for two metric throughput – continuous throughput and peak throughput. Continuous streaming is a measure of the continuous ability of the device or system over a long period of time. The peak peak indicates that the system level can provide a short period of time.

The peak capacity level is important in the VDI environment (virtual desktop infrastructure), where boot storms – when many users log into the system and start their virtual desktop at the same time – can generate large I / O requests, resulting in poor performance and increases latency if the system can not handle specs effectively.

Good channel numbers are also important for the virtual server environment when managing dynamic VM movements between datastore. Being able to measure throughput and understand peak demand is important for virtual environments.

IOPS (instantaneous input output operation) – IOPS is a measure of the number of individual reads / write that can be used by a momentary storage system. This figure is closely related but in contrast to the barriers.

In most cases, vendors will use IOPS as a measure of their product performance, but these figures need to be considered in conjunction with the size of data deductions transferred in each operation.

For example, many small requests (say, 4KB) are easier to handle than large (1MB). Also read, especially from random and uneven data sets, is generally more time-consuming than writing. Therefore, IOPS claims need to be assessed based on the number of pieces of data they handle and the types of operations they refer.

Feel free to contact E-SPIN for your storage performance monitoring requirements and solutions.