Tuesday, September 26th 2023
Solidigm Launches the D7-P5810 Ultra-Fast SLC SSD for Write-Intensive Workloads
Solidigm today announced the D7-5810, an enterprise SSD for extremely intensity write workloads. Such a drive would be capable of write endurance in the neighborhood of 50 DWPD. For reference, the company's D7-P5620, a write-centric/mixed workload drive for data-logging, and AI ingest/preparation, offers around 3 DWPD of endurance, depending on the variant; and the read-intensive drive meant for CDNs, the D5-P5336, offers around 0.5 DWPD. Use cases for the new D7-P5810 include high performance caching for flash arrays dealing with "cooler" data; high-frequency trading, and HPC.
Solidigm D7-P5810 uses SK hynix 144-layer 3D NAND flash that's made to operate in a pure SLC configuration. The drive comes in 800 GB and 1.6 TB capacities, and offers 50 DWPD over an endurance period of 5 years (4K random writes). More specifically, both models offer 73 PBW (petabytes written) of endurance. The drive comes in enterprise-relevant 15 mm-thick U.2 form-factor, with PCIe Gen 4 x4 interface, with NVMe 1.3c and NVMe MI 1.1 protocols.
In terms of manufacturer-rated performance, Solidigm is claiming up to 865,000 IOPS 4K random reads (QD256), up to 495,000 IOPS 4K random writes (QD256), and response times of 13 µs for 4K sequential writes (QD1), and 10 µs sequential read at QD1. The drive offers up to 6.4 GB/s sequential reads (128K, QD128), and up to 4 GB/s sequential writes (128K, QD128). Solidigm D7-P5810 is being offered as a persistent write buffer, and a medium tiered above capacity-focused media such as the QLC-based Solidigm D5-P5336, the pure-SLC drive stores "data about data," under a cloud-storage acceleration layer (CSAL).
Solidigm is launching the D7-P5810 series with the 800 GB model, with the 1.6 TB model joining the lineup in the first half of 2024.
A new addition to Solidigm's high-performing D7 Series, the D7-P5810 is designed for high-endurance and extreme write-intensive workloads. This ultra-fast Storage Class Memory (SCM) SSD offers up to 50 Drive Writes Per Day (DWPD) for random, 65 DWPD for sequential, and up to 2X better performance than the competition for caching, high-performance computing (HPC), data logging, journaling, and more, at less than ~20% of the cost of non-NAND SCM technologies.
SLC as a Storage Accelerator
The D7-5810 is ideally suited as a storage accelerator in front of highly dense capacity tiers (like QLC-based SSDs), adding significant benefit in the following use cases:
The D7-P5810 is available now in 800 GB (U.2 15 mm form factor). A 1.6 TB version will be available in the first half of 2024.
Source:
Solidigm
Solidigm D7-P5810 uses SK hynix 144-layer 3D NAND flash that's made to operate in a pure SLC configuration. The drive comes in 800 GB and 1.6 TB capacities, and offers 50 DWPD over an endurance period of 5 years (4K random writes). More specifically, both models offer 73 PBW (petabytes written) of endurance. The drive comes in enterprise-relevant 15 mm-thick U.2 form-factor, with PCIe Gen 4 x4 interface, with NVMe 1.3c and NVMe MI 1.1 protocols.
Solidigm is launching the D7-P5810 series with the 800 GB model, with the 1.6 TB model joining the lineup in the first half of 2024.
Solidigm Press Release
Solidigm, a leading global provider of innovative NAND flash memory solutions, is proud to announce the company's first ultra-fast, single-level cell (SLC) solid-state storage drive (SSD) for the data center market—the Solidigm D7-P5810. The D7-P5810 is a PCIe Gen 4.0 drive built on Solidigm's proven 144-layer SLC 3D NAND.A new addition to Solidigm's high-performing D7 Series, the D7-P5810 is designed for high-endurance and extreme write-intensive workloads. This ultra-fast Storage Class Memory (SCM) SSD offers up to 50 Drive Writes Per Day (DWPD) for random, 65 DWPD for sequential, and up to 2X better performance than the competition for caching, high-performance computing (HPC), data logging, journaling, and more, at less than ~20% of the cost of non-NAND SCM technologies.
SLC as a Storage Accelerator
The D7-5810 is ideally suited as a storage accelerator in front of highly dense capacity tiers (like QLC-based SSDs), adding significant benefit in the following use cases:
- Metadata/logging: placing performance-sensitive data such as metadata or logs on SLC can accelerate system performance by using SLC as a dedicated Write-Ahead Log in Ceph clusters, for example.
- Caching: SLC SSDs can act as a write buffer or cache to help remove performance bottlenecks, dramatically improve application performance, and improve TCO.
- Tiering: data is written first to the SLC SSD so commits are fast, and subsequent reads are faster. As data becomes colder, it can be aggregated, compressed, and written in bulk to underlying higher-capacity QLC drives (like Solidigm's D5-P5336) for space-efficient storage on that media. Write-shaping software such as Cloud Storage Acceleration Layer (CSAL), further delivers a solution that extends the density, TCO and sustainability value of QLC to more workloads.
The D7-P5810 is available now in 800 GB (U.2 15 mm form factor). A 1.6 TB version will be available in the first half of 2024.
Solidigm Slide Deck
42 Comments on Solidigm Launches the D7-P5810 Ultra-Fast SLC SSD for Write-Intensive Workloads
Also it is deceptive that they decided to name it P5810. The P5800X was Optane series. They are trying to convey to people that this is somehow upgrade over the P5800X although it is worse in nearly all metrics...except maybe the price (and availability) that i would hope is cheaper considering its QLC.
Of metrics, which are the most important? 4K QD1 random reads? Yes it's bad (about 75 MB/s calculated from latency) but QD1 doesn't matter in servers. Write endurance? You'll get 90,000 write cycles for the money. Real Optane has twice the DWPD rating, and the same warranty period.
Conclusion: Real Optane can't keep up, and could never keep up, given that it stayed planar (two-layer if I understand correctly) from the first day to the last day of its life.
We just need this to be released in 250 & 500 GB with m.2 interface!
Lovely news!
But you're right in that it's not marketed as pure SLC - that was just the wording in the TPU news post, and I overlooked that.
Also pSLC is considerably faster than normal QLC so its not QLC speeds, but how fast it is compared to native SLC dont know, because there hasnt really been like for like comparisons made, all SLC drives I am aware of are now really old.
As an example stacked TLC is now outperforming planar MLC, and I think has at the very least matched its erase cycle endurance.
I think the best way to market it would be 4 bit pSLC.
Its extremely rare to find SLC NAND Dies that are natevely SLC, but we do have QLC NANDs that in SLC mode can reach 100.000 PEC, for example the N48R which is micron's QLC 176-Layer 1Tb dies.its 4TB, the drive would be 3.2TB if it was QLC indeed, so if it was 3.2 TB it would have close to 37.4% of Over-provisioning
But since it's 800GB with 37.4% of OP, the real capacity is 1TB
On my DC P4600 I actually have 2TB usable, they didnt stick on 2TB and remove 30%, they added 30% on top.
What they did on this drive seems overly harsh. So 20% OP probably should be a 4.8TB drive, with 1TB usable.
Intel, Solidigm and SK Hynix are a nightmare to get information about their controllers, for real, they don't tell us anything
Of course you need to take into account that tera equals 1000^4, not 1024^4, and it's been this way since spinning rust.The calculation of total capacity is only possible when you have all the details about the die (bytes/page, pages/block, blocks/plane, planes/die), and even that can only be an approximation. There's also firmware, FTL, other metadata that take up space, then there's allowance for bad blocks (bad when factory tested), and maybe more.
Here's an example for which your database has full data: the Micron B47R die, nominally 512 Gbit, is 609 GBit raw in decimal units, or 567 Gi-bit raw in binary units. There would be more if you counted in page-level metadata (ECC etc.) but metadata can never be user data. Is that 19% OP (=609/512-1)? Or less?
But i don't count the ECC + Bad Block spares on each dies in our database.
Also the extra spare on enterprise is mostly to improve wear levelling and trim performance rather than bad block replacement, the more spare an SSD has, the more efficient it works.