If you have ever browsed the product catalogue of an industrial-grade flash storage company, you have probably stumbled across the abbreviations SLC, MLC, and TLC. If that industrial-grade flash storage company was Innodisk, you should also have spotted something called iSLC. But what do all these abbreviations mean and what is the difference between them? Let’s find out!
All About the C(ell)
Without further ado, this is what all these abbreviations stand for:
- SLC: Single-level Cell
- MLC: Multi-level Cell
- TLC: Triple-level Cell
- iSLC: Innodisk Single-level Cell
Simply put, SLC, MLC, and TLC represent three distinct types of memory cell architectures, each with their own set of strengths and weaknesses. There is also iSLC, which is a bit different, that we will look into separately.
Flash storage devices save data – ones and zeroes – in memory cells. The system determines whether a cell contains a one or a zero based on the voltage in the cell. In SLC NAND flash, the flash interprets a lower charge as a one and a higher charge as a zero. Each memory cell holds either a one or a zero, i.e., a single bit of data.
When using the single-level cell architecture, there is only one way to increase the capacity in a flash storage device: increasing the number of memory cells. However, increasing the number of memory cells only takes you so far before you start getting constrained by factors like size.
The multi-level cell (MLC) architecture tackles this challenge by allowing each cell to store twice as many charge states. In contrast to the SLC architecture, MLC flash differentiates between four different charge states, and interprets each as a 2-bit data string (11, 01, 00, 10) – allowing each cell to store twice as much information as in SLC flash. In other words, the number of cells that can store 8 GB of data when using SLC can store 16 GB when using MLC flash.
Why, then, would you stop at four different charge states? Of course, you wouldn’t. Which brings us to the triple-level cell architecture.
With the triple-level cell architecture, the flash can keep track of double the number of cell charges as with MLC, or four times as many as with SLC. The result is 50% higher storage capacity without increasing the number of cells compared to MLC memory.
A fourth up-and-coming architecture, QLC, takes the concept of increasing the number of charge states even further, bringing the total to eight states and 33% higher capacity than TLC.
At first glance, going from SLC to TLC (and beyond) seems like a no-brainer. More memory capacity with the same number of cells seems like an ideal way to keep things compact while ensuring a decreasing cost per gigabyte.
Although it is true that the technological leaps to MLC, TLC, and QLC are resulting in higher capacities and making flash storage devices cheaper than ever before, there are some drawbacks.
Benefits and Drawbacks of SLC, MLC, and TLC
If you have paid attention so far, you may have noticed that the capacity improvements with each technological leap tapers off. Going from SLC to MLC, capacity increased by a staggering 100%, but going from TLC to QLC meant a much lower 33% boost in capacity. Penta-level cell (PLC), a technology announced in 2019, would bring an even smaller 25% boost in capacity compared to QLC after doubling the number of cell charge states once again.
Declining but still impressive improvements (imagine if each new car model was 25% more fuel efficient!) is still a technological feat. However, each increase in the number of cell charge states comes with drawbacks. In fact, the much older SLC technology is still technologically superior to newer technologies like MLC and TLC in many ways.
Longer Books with Thinner Pages
You can think of a flash memory cell as a book. The SLC “book” has only two pages (one and zero), so finding the earmarked page (level of charge) takes little flipping. With TLC, meanwhile, the book stays about the same thickness, but features eight thin pages – each with three times as many characters as those in the SLC book. Since the pages are so much thinner, they tear easily, and since there is so much more information to flip through, it takes longer to find and read the earmarked page.
It is an imperfect comparison but illustrates some of the challenges when going from SLC to MLC, TLC, and beyond. With each increase in the number of cell charge states, read and write times increase. Decreasingly small voltage differences between each cell charge state also means that it takes minor wear and tear before data integrity is at risk. In SLC, meanwhile, there are only two cell charge states, making electron leakage to alter the cell charge state much less likely than in the more densely packed MLC, TLC, and so on.
Flash Types: A Comparison
Looking at the graph above, SLC looks like the clear winner for everything except capacity, whereas if capacity is the primary consideration, TLC and everything that comes after it should be the best choice. In practice, however, things are more nuanced than that.
Depending on where and how the flash storage device will be used, any one of the flash types may prove to be the most suitable one. MLC, while perhaps not “best” at anything (at least not in our simplistic illustration above), offers an excellent balance between the strengths of SLC and TLC.
Meanwhile, TLC, with its sparse number of P/E cycles, is unqualified for write-heavy workloads, for example in a surveillance setup (where SLC would be more suitable). In video hosting, for example for streaming, however, TLC may prove an excellent alternative as files are written to the drive at rare occasions before being read time after time as users stream the video files.
Beyond Flash Types: iSLC and More
While the differences between flash types are clear, the reality in the flash storage market is that there are a lot more factors that come into play when you figure out what the ideal storage device for your application is. For example, firmware and hardware enhancements can change the attributes and performance of flash devices very considerably, ensuring that they are optimized for their operating environments.
There are also pseudo-flash types such as Innodisk’s proprietary iSLC technology that combine benefits of different flash types to create an entirely new value proposition. iSLC, for instance, uses MLC flash but – just like SLC – only stores one bit of information per cell. This unique architecture allows cheaper MLC flash to gain some of the benefits of SLC flash, for example better performance and higher endurance. While, technically, still a type of MLC flash, iSLC clearly illustrates why it is unfair to treat all flash storage devices with the same flash types as fully comparable.
To make things even more complicated, there is also vertical NAND (also known as 3D NAND and BiCS Flash), which stacks memory cells on top of each other instead of storing more bits in each cell as we have looked at today. However, to keep things light, we will leave that topic for another blog post. Stay tuned!
- Flash types such as SLC, MLC, and TLC are fundamentally different architectures and store different amounts of data per cell
- Generally speaking, SLC flash performs the best while higher-density flash like TLC is cheaper and has higher capacity
- iSLC combines some of the benefits of MLC and SLC flash
- Which flash type to pick depends entirely on how and where the flash storage device will be used
- The flash type should not be the only consideration when picking an industrial-grade flash storage device