chapter 44 - flash based SSDs

Question 1

What are the three main types of flash memory cells and their trade-offs?

Answer 1

SLC (1 bit, fast & reliable), MLC (2 bits), TLC (3 bits, cheap but slow & less durable)

Question 2

What are the basic structural units in flash memory?

Answer 2

Flash is organized into pages → blocks → banks/planes. Pages are the smallest read/write unit

Question 3

What are the three fundamental operations in flash memory?

Answer 3

Read (fast), Program (1→0 only), Erase (resets entire block to 1s, slow)

Question 4

Why can’t you overwrite a programmed page in flash memory?

Answer 4

You can’t change 0s back to 1s—must erase the whole block first

Question 5

What is the role of the Flash Translation Layer (FTL)?

Answer 5

Maps logical block addresses to physical flash pages and manages read/program/erase operations

Question 6

What’s the main drawback of a direct-mapped FTL?

Answer 6

Write amplification and poor performance due to frequent block erases

Question 7

What is the key idea behind a log-structured FTL?

Answer 7

Always write to the next free page and use a mapping table to track logical→physical locations

Question 8

What causes garbage accumulation in SSDs?

Answer 8

Overwritten logical pages leave stale data behind that isn’t erased immediately

Question 9

What is garbage collection in SSDs?

Answer 9

Process of moving live data, erasing blocks with invalid data, and reusing the blocks

Question 10

What are the types of merges in hybrid FTLs?

Answer 10

Switch (fastest), Partial (some copying), Full (most expensive—copy from many blocks)

Question 11

What is wear leveling?

Answer 11

Spreads writes across flash blocks to prevent early failure of high-usage blocks

Question 12

Difference between passive and active wear leveling?

Answer 12

Passive uses natural spread via logging

active relocates static data from low-wear blocks.