Someone recently asked The SSD Guy if there is a way to determine whether an SSD is SLC, MLC, eMLC or TLC.
I found it a little odd to be asked this, since most vendors tell what kind of flash they use in an SSD’s specifications, especially if it’s SLC.
Not finding it there then the next thing I would look at is the price. Raw SLC NAND flash now sells for about 6-10 times as much as its MLC counterpart, so an SSD with a price of around $1/GB is likely to be MLC and one that sells for around $10/GB is probably SLC.
TLC SSDs are really rare. There is the 840 from Samsung and the Ultra II and X300 from SanDisk. It’s also pretty rare to find an SSD that is based on eMLC, because eMLC achieves its higher endurance by slowing the part down. There are other, better ways to extend endurance.
Now that you know all this, please note that the kind of flash used to produce the SSD is probably unimportant. Certain MLC SSDs outperform some SLC SSDs, both in speed and in endurance – it all has to do with the quality of the controller. Some useful comparisons of SLC and MLC SSD performance are in the post: “Not All SSDs are Created Equal“. There’s a thorough explanation of controller techniques in The SSD Guy’s series on SSD controllers.
Furthermore, NAND chips allow controllers to change flash pages inside the chip from TLC to MLC to SLC as needed, so some SSDs actually internally manage hot data to SLC and cooler data to MLC or TLC.
For those that have read this far, I would recommend changing your focus away from questions of SLC vs. MLC vs. TLC and focus instead upon the SSD’s published specifications (like speed and endurance) and on its SMART attributes. The SMART attributes will always tell you how your SSD is wearing, and the specifications will tell you how it is expected to perform.
If you don’t trust the manufacturer’s performance specifications, then I highly recommend using the SNIA performance test specification, which gives unbiased performance results for any SSD.
Erik Logan of Pogo Linux sent me a link to an amusing & informative video he and his company put together called A Brief History of SSDs. In the video Erik (pictured) tells of Pogo Linux’ experience with SSDs.
The company has a lot of hands-on SSD experience: Pogo Linux ships servers and storage and has ramped SSD shipments (as a percent of all drives) from single digits three years ago to the point where SSDs now account for 31% of all drives they ship. Erik shared with me that: “Sorting through the Continue reading
Earlier this month Western Digital’s HGST division invited The SSD Guy to a launch of a number of products. On the HDD side there were:
- 6TB air HDD, HGST’s last air-filled enterprise HDD
- 8TB helium HDD, an incremental upgrade of last year’s 6TB helium HDD
- 10TB shingled helium HDD (pictured)
I view these as very solid evidence that HDD costs will continue to stay an order of magnitude cheaper than SSD costs, thwarting the price-per-gigabyte crossover that others have been predicting for years.
In fact, since my last post on the price crossover in 2011, very little has changed.
It’s safe to assume that the HDD industry will Continue reading
Someone recently asked The SSD Guy to guess what would be the largest amount of flash that could be fit into an SSD’s case. This sounded like a fun problem, so I did a “Back-of-the-Envelope” estimate to try and figure it out.
First of all, I would judge by this post’s picture that you could get no more than 20 chip packages (4 x 5) on one side of a PC board for a 2.5″ SSD. That’s probably an optimistic estimate.
If you ignore the controller that would allow you to squeeze 40 packages onto a single circuit board.
Certain high-capacity SSDs use a “Butterfly” design to fit three circuit boards into a single 2.5″ HDD housing. With three 40-package circuit boards you could fit 120 chip packages into the 2.5″ HDD housing.
Today’s densest flash chip stores 128 gigabits or 16 gigabytes. Samsung and SanDisk can stack 16 of these chips within a single package, making a 16 x 16 gigabyte or 256 gigabyte package. SanDisk just announced a 512 gigabyte SD Card that doubles Continue reading
Tom Coughlin and I have just released a new report that helps shed a lot of light on a pretty challenging subject: We asked nearly 200 IT managers to tell us how much storage performance their systems require. They provided candid replies about their IOPS, latency, and capacity needs for a number of leading applications.
This study breaks the market into 23 application types, and provides an explanation of each along with forecasts by major application category.
Virtualized systems will drive the greatest 5-year average unit shipment growth, at 85%, although the data center will retain its leadership in enterprise SSD consumption. Overall enterprise SSD unit shipments will grow at an annual average of 32% through 2018.
Since SSD prices are cost-based, with roughly 80% of the cost coming from flash chips, NAND flash price swings will cause Continue reading
On Wednesday OCZ announced that its bank accounts had been seized by one of its creditors and that the company would file for bankruptcy, but it did not commit on which of two courses of action it would take:
- To file for bankruptcy and sell itself as an ongoing business to Toshiba
- To file for bankruptcy and liquidate
If the company is sold to Toshiba the bankruptcy court will require an auction to be held to assure that the price that Toshiba pays is the best price that the company can get. This means that there is still the possibility of another company actually acquiring OCZ. Although Seagate was rumored to be interested there are certainly others who are also preparing bids.
OCZ has good technology and a loyal retail customer base, but one year ago Continue reading
The results of the Storage Performance Council’s SPC-1 report, show Kaminario surpassing last year’s record performance by 20k IOPS.
Interestingly enough Kaminario set the 2012 record using DRAM while this year the company was able to do it with its fourth-generation all-flash K2.
Why is the SPC1 test so highly respected in Continue reading
Violin Memory and Microsoft have jointly announced a novel way of harnessing the power of Windows Server software. Violin will be shipping its memory arrays with a special version of Windows Server 2012 R2 pre-installed on the embedded server that manages the internal operations of Violin’s all-flash array.
Violin explains that native support of specially-optimized versions of Windows Server and System Center that have been tuned for an all-memory array will provide improved performance and economics for large-scale enterprise cloud deployments.
The system can internally run Continue reading