SSDs and Flash Memory
How does memory that never forgets, even without power, work?
Summary
SSDs, phones and memory cards keep their data in flash memory. Each flash cell is a transistor with a "floating gate" buried between two insulating layers. When writing, a high voltage pushes electrons through the thin insulator and traps them in the floating gate; there they stay for years without power. Flash has one oddity: a page can only be written while it is empty, and erasing is done for a whole block. So when an SSD changes a file it writes the new copy somewhere else and cleans out old blocks every so often.
In depth
Flash memory was invented in 1980 by Fujio Masuoka's team at Toshiba. The amount of charge in a cell's floating gate sets the threshold voltage needed to switch the transistor on; reading means comparing this threshold with read limits. Writing and erasing push electrons through an oxide a few nanometres thick by Fowler-Nordheim tunnelling, and every pass wears the oxide a little. Storing 2 (MLC), 3 (TLC) or 4 (QLC) bits per cell increases capacity but narrows the margins between levels and shortens the lifetime. The controller in an SSD maps logical addresses to physical pages, cleans out stale pages with garbage collection, wears blocks evenly and repairs errors with error-correcting codes. The TRIM command tells the SSD about deleted files and so prevents needless writes.
Rate SSDs and Flash Memory
What did you think of this content? Click the stars to leave your rating.
Leave a Comment and Rating
Sign in to comment and rate. Membership is free.