NVMe NAND based SSDs vs. SAS SSDs? How do you think this will progress in DC over the next 5 years?

cbn

Lifer
Mar 27, 2009
12,968
221
106
DC= data center.

Some factors to consider:

1. Optane (and other non-volatile memory) DIMMs

2. Optane NVMe (most likely MLC based*) ---> https://www.anandtech.com/show/1120...-dive-into-3d-xpoint-enterprise-performance/2

*Especially if future generations of Optane lower latency therefore making the multi-bit per cell Optane of the future equal in latency to the SLC Optane of today.

3D XPoint is currently storing just one bit per memory cell while today's NAND flash is mostly storing two or three bits per cell. Intel has indicated that the technology they are using, with sufficient R&D, can support more bits per cell to help raise density.

3. Yangtze Memory Xtacking NAND (high throughput per GB) ---> http://www.ymtc.com/index.php?s=/cms/172.html

“At present, the world’s highest 3D NAND I/O speed is targeting 1.4Gbps while the majority of the industry is offering NAND I/O at 1.0Gbps or below. With our XtackingTM technology, it is possible for NAND I/O speed to reach up to 3.0Gbps, similar to I/O speed of DRAM DDR4. This is going to be a game changer in the NAND industry,” said Simon Yang, CEO at YMTC.

4. Samsung Z-NAND --> https://www.anandtech.com/show/1170...summit-96layer-vnand-mlc-znand-new-interfaces

5. Toshiba low latency NAND (pretty interesting stuff---> https://www.anandtech.com/show/13183/toshiba-announces-xlflash-lowlatency-3d-nand )

IMGP1992_678x452.jpg



6. Segate Multi-actuator SAS hard drives (spinning at low rpm with a high actuator count. Also potentially with Dual Pillar)
 

nosirrahx

Senior member
Mar 24, 2018
304
75
101
Its hard to say but it will be in the best interest of all involved if future solutions can be used within existing systems.

An amazing new technology that requires an entire new server with new ports is a hard sell.

I can see BBU volatile cache going away completely, this is the one DC technology that needs to become obsolete.

Multi-actuator feels late to the party, it would have been a killer product maybe 5 years ago.

I see the issue with tiny random files and massive files benefiting from completely different solutions being resolved through the creation of hybrid products. Monolithic products based on new age storage technology get crazy expensive at large capacities. It would be way more cost effective to build AIO devices that use both older and cheaper NAND for sequential data and mirror the latency sensitive data to new age technology.
 

cbn

Lifer
Mar 27, 2009
12,968
221
106
Multi-actuator feels late to the party, it would have been a killer product maybe 5 years ago.

I don't think it is late to the party myself, but I do wonder why it didn't show up sooner?

One guess I have is Seagate is seeing/predicting the need for more performance per watt Sequential Read* (in addition to the touted increased IOPs). (Example: In this greybeards podcast even Western Digital admits average block sizes are growing in this era of rich data. (reference 28.30 (and 28:40 ) to 29:26 in the pod cast. So throughput, much more than IOPs (<---This mentioned at 28:47 to 28:50 by one of the greybeards)). Of course, RAID rebuild times is another thing to consider (boosting Sequential helps this) and I wonder how much it helps file transfer within a server system like this one which uses the Exalytics in-memory machine).

*and performance per dollar Sequential Read.
 
Last edited:

nosirrahx

Senior member
Mar 24, 2018
304
75
101
I don't think it is late to the party myself, but I do wonder why it didn't show up sooner?

The main benefit I see here is combining the performance of the 15K RPM drives with the capacity of the 7200RPM drives.

The biggest 15K drive I have ever seen was pretty small.
 

cbn

Lifer
Mar 27, 2009
12,968
221
106
The biggest 15K drive I have ever seen was pretty small.

Yep, the current ones from Seagate are only 300GB, 600GB and 900GB. This using 300GB platters.

P.S. Below is a size comparison of two 2.5" hard drives (A 5400 rpm 500GB SSHD and a 15,000 rpm 600GB SAS drive) from this tear down article (Picture gallery with 40 images here) :

(Notice how small the platters are from the 15K 2.5" drive....I believe they are only 1.8")
ImageResizer.ashx



ImageResizer.ashx



It takes 3 of the smaller 200GB platters (from the 15K SAS drive) to exceed the capacity that one 500GB platter (from the SSHD).

The following reasons given on this page explain why:

The Laptop drive spins at 5400RPM and contains 500GB of storage on one platter, the Enterprise drive contains 600GB storage over 3 platters. The platters are smaller on the Enterprise drive because it spins nearly 3 times the speed at 15,000RPM. Smaller diameter disks are less resistance to vibration. Vibration is a very important factor in server drives because if you have too much vibration induced by so many drives squeezed into the server chassis; vibrations can magnitude to the point that the head can no longer stay in line to the data tracks. By lowering the aerial density on the platters, a wider data tracks can be made, this makes it better for the head to stay on track at higher vibrations and higher disk RPM speeds, it also increases the data integrity required in mission critical applications that can be found in servers. The enterprise drive also features a much heavier/studier base plate. This also helps against drive vibration and also heat dissipation where multiple drives are mounted close together.

So the capacity is low for two reasons:

1. Small diameter platter. (This keeps the seek time low for the actuator and energy consumption relatively low despite the high RPM. In fact, the current 15K 900GB model (with three platters) has an operating power of only 7.6W....the single platter 300GB model comes in at 6.9W. This, contrast, to the 8W of a 7200 rpm 3.5" 2TB Barracuda ST2000DM006).

2. Wide tracks to better handle vibration in a rack.