Kingston DC1500M 1.92TB Data Center Enterprise SSD Review

TweakTown's Rating: 93%

The Bottom Line

The DC1500M delivers the goods over a broad range of datacenter applications, making it one of the best Gen3 options on the market.

Introduction and Drive Details

Kingston Digital is now shipping the DC1500M enterprise SSD for datacenter applications. The DC1500M supports up to 64 namespaces, making it ideal for virtualized environments and applications such as web hosting. The drive has at its heart a Silicon Motion Inc. 16-channel SM2270 controller. SMI's SM2270 controller has built a well-deserved reputation for delivering consistent, predictable performance under a broad range of datacenter workloads.

For flash, Kingston is using BiCS4 96Layer TLC. We like this choice as it has proven to deliver random read performance superior to other 96Layer flash choices. A 16-channel Silicon Motion SM2270 controller managing a BiCS4 flash array should be a recipe for high QoS (Quality of Service), which is exactly what Kingston is claiming the DC1500M is all about. Let's get Kingston's newest datacenter SSD on the bench and see if the results match the claims.

Quick Specs/Comparison Products

29

Kingston DC1500M 1.92TB Enterprise PCIe Gen3 x4 NVMe SSD

29

This side of the drive features a familiar-looking manufacturer label.

29

The bottom of the drive's enclosure is a heavy-duty cast aluminum piece that serves as a heat sink.

Specifications and Features

29

Kingston's PCIe Gen3 x4 NVMe SSD is a 2.5" x 15mm U.2 (SFF-8639) Enterprise SSD available in capacities ranging from 960GB up to 7.68TB. Features include a PCIe Gen3 x4 U.2 interface, SMI NVMe controller, BiCS4 TLC flash, up to 64 namespaces, power loss protection, end-to-end data protection featuring advanced LDPC bit correction on memories in the data path for protection at every layer. Endurance is good at 1 drive write per day for 5-years.

Current online pricing finds the 1.92TB DC1500M retailing for approximately $484.

Enterprise Testing Methodology

TweakTown strictly adheres to industry-accepted Enterprise Solid State Storage testing procedures. Each test we perform repeats the same sequence of the following four steps:

1. Secure Erase SSD
2. Write entire capacity of SSD a minimum of 2x with 128KB sequential write data, seamlessly transition to next step
3. Precondition SSD at maximum QD measured (QD32 for SATA, QD256 for PCIe) with the test specific workload for a sufficient amount of time to reach a constant steady-state, seamlessly transition to next step
4. Run test specific workload for 5-minutes at each measured Queue Depth, record results

Benchmarks - Random and Sequential Performance

4K Random Write/Read

29

We precondition the drive for 16,000 seconds, receiving performance data every second. We plot this data to observe the test subject's descent into steady-state.

Steady-state is achieved at 6,000 seconds of preconditioning. The average steady-state write performance at QD256 is approximately 240K IOPS. The relatively tight pattern with virtually no outliers indicates a high QoS.

29

29

We have no problem sustaining 240,000 4K random write IOPS which is 20K better than Kingston's stated sustained random write spec. Kingston's newest delivers the second highest write performance for any flash-based SSD we've tested to date. When we factor in its outstanding low queue depth performance, we are of the opinion that the DC1500M delivers the best-in-class 4K random write performance. Outstanding

29

29

Kingston specs the DC1500M 1.92TB for up to 510,000 random read IOPS, and again we exceed factory specs by roughly 20K IOPS. Overall, the DC1500M is delivering average 4K random read performance.

8K Random Write/Read

29

We precondition the drive for 16,000 seconds, receiving performance data every second. We plot this data to observe the test subject's descent into steady-state.

Steady-state is achieved at 7,000 seconds of preconditioning. The average steady-state write performance at QD256 is approximately 112K IOPS.

29

29

We expect 8K random to track exactly the same as 4K random, just at a 50% lower rate. This is exactly what we get. Again, we will point out the DC1500M is delivering the best performance where it matters the most, QD1, for any flash-based Gen3 SSD we've tested. Excellent.

29

29

Here we expect to see our chart tracking the same as our 4K chart with somewhat lower performance. Results come in better than expected, as we find the DC1500M delivering better than average performance at QD1-QD32. Nice.

128K Sequential Write/Read

29

We precondition the drive for 6,500 seconds, receiving performance data every second. Steady-state for this test was already achieved when we filled the drive with 128K seq data. A whole string of outliers going in the wrong direction indicates poor QoS when writing data sequentially.

The average steady-state sequential write performance at QD256 is approximately 2,200 MB/s.

29

29

Our testing indicates factory sequential write specifications to be overstated by almost 500 MB/s. However, when we step back and look at its performance at QD4 and QD8, we find it's the best we've seen from a 2TB class PCIe Gen3 enterprise SSD to date. We think that the factory specs of 2,700 MB/s for the 1.92TB model might be a typo.

29

29

It looks like the DC1500M is trying to make up for underperforming vs. stated sequential write spec by exceeding stated spec by roughly 400 MB/s when reading sequential data. That's a nice trade-off that we will gladly take any day as read performance is more important than write for the majority of applications. Additionally, and worth noting is the fact that at QD8-16, the DC1500M is the best we've tested.

Benchmarks - Server Workloads

Email Server

An Email Server workload is a demanding 8K test with a 50 percent R/W distribution. This application gives a good indication of how well a drive will perform in a write-heavy workload environment.

29

We precondition the drive for 16,000 seconds, receiving performance data every second. We plot this data to observe the test subject's descent into steady-state.

Steady-State is achieved at approximately 14,000 seconds of preconditioning. The average steady-state workload performance at QD256 is approximately 90K IOPS.

The small number of outliers indicates average to above average QoS.

29

29

The typical operating region for most enterprise workloads is in the QD1-QD16 range. The DC1500M delivers the best Gen3 performance we've seen here at QD1-8. This is where we want to see exceptional performance, and this is right where the DC1500M is serving it up in a big way.

OLTP/Database Server

An Online Transaction Processing (OLTP) / Database workload is a demanding 8K test with a 66/33 percent R/W distribution. OLTP is the online processing of financial transactions and high-frequency trading.

29

We precondition the drive for 16,000 seconds, receiving performance data every second. We plot this data to observe the test subject's descent into steady-state.

Steady-state is achieved at 15,000 seconds of preconditioning. The average steady-state workload performance at QD256 is approximately 112K IOPS.

The small number of outliers indicates average to above average QoS.

29

29

Again, the DC1500M shows us best in class performance at low queue depths. Performance that matters.

Web Server

A Web Server workload is a pure random read test with a wide range of file sizes, ranging from 512B to 512KB at varying percentage rates per file size.

29

We precondition the drive for 16,000 seconds, receiving performance data every second. We plot this data to observe the test subject's descent into steady-state.

We precondition for this test with an inverted (all-write) workload, so no relevant information can be gleaned from this chart.

29

29

More good stuff from the DC1500M. We find it is essentially tied with the DC-P4510 for the best Gen3 performance at queue depths up to 8.

Final Thoughts

Kingston's DC1500M displayed its prowess over a broad range of datacenter use case scenarios we tested it against today. The drive impressed with its low queue depth performance and demonstrated that it is a force to be reconned with as it delivered some of the best PCIe Gen3 sustained performances we've seen to date.

The DC1500M 1.92TB impressed us across the board, with the lone exception of not delivering sequential write performance that matches stated performance. This isn't a big deal because it is delivering exactly as expected for a 2TB class PCIe Gen3 datacenter SSD, maybe even a little better.

29

Kingston's newest datacenter SSD delivered performance where it matters, some of it we could even call best in class, and has earned one of TweakTown's highest awards and a hearty recommendation.

Pros:

• Best Form Factor
• Low Cost
• Multiple Namespaces
• Density

Cons:

• Seq Write