Testing and comparing the performance of your Local and Shared Storage

Fill in the Form to Continue
img

Host:
Dmytro Khomenko, Pre-sales Engineer, StarWind

Duration: 25:03

PUBLISHED/UPDATED: June 22nd, 2017

Key points of the webinar:

  • Storage performance depends on the software and hardware stack utilized
  • IOPS, latency and network bandwidth – main storage performance indicators
  • StarWind Storage Test tool – simple and powerful instrument for storage benchmarking
  • StarWind shared storage provides higher IOPS and lower latency comparing to local storage

When testing the storage performance, it must be noted that it highly depends on the type of software and hardware stack that is being used. In general terms, the performance of a system has four major factors: memory, processing power, network bandwidth, and disks (IOPS and capacity).

The most important aspect to be considered when testing a highly available storage device is that the HA performance depends directly on the network and disks, therefore it is essential to ensure that the SAN is configured properly before it is put in use.

There are several other things that must be considered when testing performance, and storage type is among the first.

Random performance is one of the key limitations of HDDs and something that SSDs excel at. HDD manufacturers tend to show average latency statistics as an indicator of random performance, while SSD vendors provide 4KB random IOPS figures.

Unlike HDDs, flash-based storage products display significant performance differentiation among different models. It should be noted that SSD performance changes over time. There is a period between the “fresh-out-of-the-box” and the “steady state” period where there is a “transition” period.

When talking about RAID, the speed of a RAID array is only as fast as the slowest member, thus an errant I/O from one drive slows the speed of the entire array in most scenarios. Writing data to a RAID set or flash drive is often substantially slower than reading that data back. Therefore, most benchmarks include a mix of READ and WRITE operations, often skewed toward the former.

To calculate the actual IOPS for a given disk, the following information is required: average latency and average seek time.

Although latency is the most important measurement for most of the workloads, in reality, every storage device provides a wide distribution of variable latency while under load, even when aggregated into RAID arrays. The problem is that popular performance testing tools need a lot of scripting and automation to simulate that kind of environment.

To perform the testing, the StarWind Storage Test has been used. It is a powerful tool designed to provide a simple means of assessing the read and write performance of SSD or HDD and comparing it with highly available fault-tolerant storage. The test results show that the StarWind HA disk provides the improvement of IOPS in both random and sequential testing patterns and shows lower latency.

Power Your Business With StarWind!
Book Free Online Demo:

Thank you for your application!
We will reply to you shortly.