Flash storage

Storage in flash is frequently discussed in the support channel. Toradex invests significant resources to make the NAND flash board as reliable as is possible. But, it’s important to know the basics of the storage device that is used. One of the primary aspects to be aware of is that, if your storage gets worn out it is possible to damage your storage device by writing too much data to the Flash memory controller device. In this article we’ll give you an outline of the potential problems flash storage could face. Let’s begin with a quick background on the technology.

Types of Flash: Raw Flash vs Managed Flash

Presently, Toradex computer modules used NAND, NOR, as well as flash eMMC.

NOR as well as NAND are storage devices that can be used as raw. NAND however, is only accessible in pages. A few bits of the page could be incorrect and require correction through an error correction mechanism.

With NAND and flash, on the other hand, the device driver and OS are in charge of handling these issues.

Evolution of NAND Flash: From SLC to MLC

The bit density of NAND flash has increased in the course of time. Initial NAND flash devices were single Level Cell (SLC) flash. This means that every flash cell is able to store just one bit. In the case of multi Level Cell (MLC), flash cells can hold more than two bits per cell, which means that the density of bits increases. This sounds fantastic, however it’s not all that great. MLC there are some drawbacks too as With MLC NAND, you also get an increased rate of error in bit and shorter endurance. All eMMC use MLC NAND. Certain eMMC devices let you change to the”pseudo-SLC” (PSLC) mode for a portion or (or) all storage.

Endurance: Limited number of erase cycles

As we’ve already said among the most important aspects to be aware of about flash technology in the devices we use is that you are able to create and erase data for a certain number of times. Writing large amounts of data into the flash device isn’t an ideal idea! According to the above table based what type of flash you are using you can have between 10 and 100K erase cycles before data is damaged or disappears. The phrase “erase cycles” is irritating. One of the limitations that flash storage has is that it can’t be written without having been erased prior to. Furthermore, this can’t be accomplished at the bit level however, it can only be done at larger chunks known as erase blocks. In the worst-case scenario this would mean that when you just desire to write a single byte, you may need to erase and write a whole block. The size of the block, erase could reach up to the size of 512 bytes. The effect of erasing/ the writing you need is referred to as writing amplifying. There may be additional write operations required in the filesystem flash. If you are trying to determine the lifespan that the storage system in flash in the device you are embedded on, make sure to consider this.

The flash memory of NOR

NOR and NAND make use of distinct logic gates — the primary electronic circuit’s building blocks inside each memory cell in order to store the data.  

NAND flash contrasts with. NAND flash

The NOR flash memory is quicker to read than flash NAND, however it’s also more costly and the process takes longer for erasing and writing in new data. NAND is more efficient in capacity memory than the NOR.

However, the older more traditional, parallel NOR flash technology can support the one-byte access to random data. This allows machine-specific instructions to be read and executed direct from the flash similar to how computers typically retrieve their instructions directly from the main memory. It is however, NOR must write in larger amounts of data per minute that NAND. 

The NOR flash technology is more costly to manufacture and comes with a higher price for each millimeter than NAND.