How do solid state drives (SSD) store and retrieve data?
In: Technology
On a very basic level, an SSD is like a large plastic sheet covered with a grid of those bumps that you see on soda-cup lids. To write a file to the disk, you pop and unpop the bumps to represent the file in terms of 1s and 0s. To read from the disk, you run your finger across them and feel the pattern of popped and unpopped bumps.
These bumps are called flash memory cells, and there can be billions or trillions of them packed in several tens of layers on a silicon chip. They consist of microscopic transistors and capacitors, and they’re ‘popped’ and ‘unpopped’ by applying a positive or negative voltage across them, which either stores a tiny amount of electric charge or releases it. To check the state of one of these cells, you can measure the voltage across it to see whether there is charge stored in that cell. A controller chip on the SSD receives read and write commands from the host computer and applies the necessary voltages to the right rows and columns of cells to write and read blocks of several thousand cells, called ‘pages’. The exact format of the data on the disk is entirely up to the controller and may not be consistent across different models of SSD’s — so long as the controller implements a common protocol (SATA or NVMe, for example), the computer doesn’t care.
These cells aren’t perfect, and repeatedly writing and erasing a cell causes it to wear out and lose its ability to store charge. When this happens, the controller can no longer read from those cells, causing data loss. To combat flash memory wear, an SSD controller is smart enough to spread out the reads and writes across the many billions of cells at its disposal. This process is known as ‘wear-leveling’. Even with this algorithm in action, a disk has a limited amount of read-write cycles before the probability of a data loss incident becomes more than negligible. This quantity is usually expressed in write cycles, where one write cycle consists of using every cell on the disk once. Because of this, larger-capacity SSD’s take longer to fail on average.
Latest Answers