non-volatile memory

Think of the difference between writing on a whiteboard and writing in a notebook. The whiteboard (DRAM and SRAM) is fast to write and erase, but wipe the room's power and everything vanishes. The notebook keeps your words even after you close it and walk away. Non-volatile memory is the notebook of the computer world: storage that retains its contents with no power applied, so your data survives a shutdown or a crash.

Concretely, non-volatile memory stores bits in a way that does not depend on a continuously supplied charge. NAND flash, the dominant kind, traps electrons on a floating gate inside each transistor; the trapped charge persists for years and is read back as a 0 or 1. This is the technology inside SSDs and memory cards. Emerging non-volatile memories store bits differently: phase-change memory (PCM) switches a material between crystalline and amorphous states; MRAM uses magnetic orientation; ReRAM changes a material's electrical resistance. All of them keep their state without power, and the emerging ones aim to be far faster than flash, closer to DRAM speeds.

Why it matters and the honest caveats: non-volatility is hugely valuable — it is why your files survive a reboot — but today's non-volatile technologies have real costs that keep them out of the role of main memory. NAND flash is much slower than DRAM, is written in large blocks (not single bytes), and wears out after a limited number of writes (hence wear-leveling). The emerging persistent memories are faster but still face cost, density, endurance, or write-energy hurdles. The exciting prospect they raise is storage-class memory: a tier fast enough to act like memory yet persistent like storage, blurring the old line between the two.