Most of the test system hardware for NAS testing isn’t an important factor when it comes to benchmarks, most of it comes down to the network infrastructure and drives. I am using 10 GBASE equipment which means that I should max out any Gigabit Ethernet connections with ease and I got a second port to aggregate with, in case one isn’t enough. The benchmarks will be performed with Seagate IronWolf 10TB NAS drives and through a 10GBASE managed switch which should give us some great benchmark results.
The NAS is filled with HDDs and all available drive modes will be tested. Which are available will depend on the tested model. However, all available modes will be tried and benchmarked. That includes RAID 0/1/5/6/10, JBOD, and single drive setups, but also SHR1 and SHR2 where available. I will also test different file systems such as EXT4, ZFS, and Btrfs as well as folder and volume encryption whenever they’re available.
We use the best benchmark result from each test, and each benchmark is run five times five times. In the event of large differences, multiple charts will be created to show the different performance results.
The NAS connects through a D-Link DXS-1210-12TC managed 10GBase switch which connects to the rest of our network. All client systems use 10GbE adapters of varying origin, but the main NAS and network test system has a D-Link DXE-820T dual 10GbE network interface card.
With a device of this sort having so many different applications, Intel’s NASPT software covers all the bases and also gives us a set of results that we will be able to utilise and compare those against other similar systems.
Intel NASPT (Network Attached Storage Performance Toolkit) performs its test by transferring varying sizes and quantities of data to and from the device based on twelve different scenarios.
Intel’s NASPT software does require us to drop the memory down to 2GB, as using any more would lead to data caching and skew the results. But it’s a great and free app, allowing you to run your own comparisons.
Normally, all benchmarks would be run with the Btrfs file system when available. That isn’t the case here as the feature was rolled out after my benchmarks all where done with EXT4 and compiled into the result charts. I tested Btrfs on a single volume after the latest patch and added those results as a seperate item.
Expansive Dimensions, Big Backdrop: complete background coverage, more room to move freely, fit more people…
80 PLUS GOLD CERTIFIED – An efficient power supply reduces energy consumption and heat load…
A NAS is a common bit of kit for a small office as it provides…
Cooler Master has just revealed their new series of gaming chairs, the CALIBER X2C and…
Thermaltake already has many of my favourite CPU coolers on the market right now, especially…
In July last year, Netflix officially confirmed that it had ended the option for new…