Looking to replace SSD drives in your data center? Give the Hynix Gold S31 SSD a spin.
You probably have a lot of servers powering your data center. That means you also have a lot of drives to deal with. Drives fail. Drives must be replaced. Some of those servers probably contain old school platter drives that are probably on their last legs. The likelihood of you wanting to replace those disks with similar mechanical drives is doubtful. Instead, you’ll want to take a more modern approach and install solid state drives (SSDs). SSDs have been proven superior to platter drives–they’re more reliable (they don’t have moving parts), significantly faster, use less energy, and take up less space.
The one disadvantage of SSDs is price. Whereas a TB SSD drive will cost you anywhere from $100 to $300 USD, you can purchase an 8TB platter drive for $200.00. That’s a significant difference in cost. Even still, SSD prices are dropping and a TB drive for around $120.00 is common. So when Hynix reached out to me to plug their Gold S31 SSD SATA drive ($118.00 on Amazon), I figured it couldn’t hurt to compare it to my current Samsung SSD drives in my System76 Thelio. I’d never heard of the company, so I assumed them to be selling typical generic hardware that couldn’t stand up to the tried and true drives I’ve been using for some time.
I cases like this, I always welcome being wrong.
After inserting and formatting the drive, the first thing to be done is run a comparison benchmark test between the Hynix Gold S31 SSD and a Samsung 860 EVO. The hosting operating system was Pop!_OS 19.10 and the benchmarking tool was GNOME Disks.
SEE: Flash storage: A guide for IT pros (TechRepublic Premium)
After running both read and read/write tests, I was surprised at how similar the results were. Running a read test of 10,000 samples, the Hynix had an average read rate of 555.7 MB/s, with an average access time of 0.04 msec (Figure A).
The Samsung results were similar, with a 558.5 MB/s average read rate and a 0.05 msec average access time (Figure B).
It should come as no surprise that the read/write testing results came out similar as well. For the Hynix, the results were a read rate of 554.7 MB/s on 100 samples, an average write rate of 497.8 MB/s, and an average access time of 0.04 msec (Figure C).
The Samsung EVO read/write test came out with a 554.1 MB/s average read rate, 498.5 MB/s average write rate, and an average access time of 0.11 msec (Figure D).
Although the difference between the Samsung EVO and the Hynix Gold S31 are quite negligible, when you’re talking at scale with massive amounts of I/O, those numbers quickly add up.
Data copy test
I also did a data copy test with the Hynix. Copying a folder that contained 98.9 GB of data, took approximately two minutes from the Samsung to the Hynix. From the Hynix to the Samsung, the time was closer to three minutes. Granted, these tests were executed at different times, so there could have been a slightly heavier load on the system during the Hynix to Samsung test. Nevertheless, nearly 100 GB of data copied in two minutes is solid performance.
The spec sheet on the S31 looks like:
- Capacity: 250GB/500GB/1TB
- Interface: SATA III 6Gb/s
- Form factor: 2.5″
- Sequential Read Speed: Up to 560MB/s
- Sequential Write Speed: Up to 525MB/s
- TBW (Terabytes Written): Up to 600TBW
- Controller: SK Hynix Controller
Although my test drive didn’t quite live up to the max write speed, it certainly came close to the max read speed. Even so, 498.5 MB/s is still pretty fast.
Who is this drive for?
Whether you’re a gamer, a developer, or an admin looking to replace drives in your data center servers, the Hynix might be the solution. The company has been around since 1983, starting out as Hyundai Electric Industry, then shifting to Hynix Semiconductors in 2001, and finally making the move to SK Hynix in 2012–so they know technology. Give one of these babies a try and see if they don’t live up to–or surpass–your expectations.