One Little Bit: Hall-Effect Core Memory circuits that remember their state.
Core memory has existed the 1950s, it stores a value of a 1 or 0 in a ferrite loop — a tiny electromagnet core. To write a 1 value, you push current and magnetize the core one way, to write a 0, you push current the opposite way. Compared to modern memory systems, it’s slow and cumbersome, one problem is that reading the data destroys the data. You push a small current through the core, and you check to see if the output is low or high. Unfortunately, this overwrites whatever value was there. So, if you plan on re-using that value, you need to restore it just after you read it.
Core Memory: Indestructible, but erased every time it’s read
The really interesting aspect of this memory is that it is not only non-volatile, but nearly indestructible. Even after the Challenger Shuttle disaster, the memory in the flight computers could be re-read, since those tiny magnets still held their values. This allowed the state of the computers to be re-read at the moment the systems lost power. These non-volatile cores could conceivably store data forever, as long as they are protected from stray magnetic fields. They can be written and read indefinitely, and will never break.
SD Cards: Dense, reliable, but will eventually fail
