Solid State Drives (SSD) 101
In part one of this series we covered the internals of HDDs and some basic techniques manufacturers use to increase performance. In part two we are going into a deep dive of Solid State Drives (SSD), how they work, and some caveats.
Solid State Drives (SSD)
The solid-state drives (SSDs) have a simple unit where one or more bits are stored: the “NAND Flash Cell”. So, this should be easy! Right! Lets see.
The SSDs are constructed like a Lego where the smaller piece is the “Flash Cell”. We aggregate multiple “Flash Cells” into a “4KB Page”. The amount of “Flash Cells” in a “Page” depends on the amount of bits the “Flash Cell” can manage. Now, here is the first caveat, a “Page” is the minimum writable unit in SSD. Even if you need to write a single bit, you would have to write an entire “4KB Page”.
We take “Pages” and group them into “512KB Blocks”. Here comes the second caveat, “Blocks” are the minimum erasable unit in SSD.
This causes a phenomenon known as the “write amplification” effect in SSD. If you need to erase a single bit, you need to modify and entire “Page” (the minimum writable unit), but you can only erase a “Block”. The drive needs to read the “128 Pages” that made the “Block”. Next it erases the “Block”, then write back the 127 unmodified “Pages” plus the 1 modified “Page”. Read More »