What it is and the problem it solves
DRAM is a volatile semiconductor memory technology that trades storage stability for density. It solves the problem of building cheap, high-capacity RAM when transistors were expensive and silicon area was scarce.
How it works
Each DRAM cell stores one bit as charge on a capacitor, controlled by a single MOS transistor. Reading the cell discharges it slightly, so the sense amplifier must restore the charge — which doubles as refresh. Rows are refreshed every 64 ms or less, per JEDEC.
What works
The single-transistor, single-capacitor cell works. The Intel 1103 worked in 1970. Multiplexed addressing on the Mostek MK4096 worked in 1973. Sense-amplifier read-refresh works — reading a row refreshes all cells in that row.
What does not
It does not retain data without power. It does not avoid refresh overhead. It does not eliminate access latency: row activation and precharge add delay that SRAM avoids. It does not scale linearly — refresh cycles consume bandwidth and complicate timing.
What it changes
It changed memory economics. One transistor per bit cut cost and size versus six-transistor SRAM. It enabled kilobit-scale memory chips by 1970, then megabit and gigabit densities. It made volatile main memory affordable and scalable — at the cost of constant housekeeping.
Is it worth your time
Yes, if you work with memory systems, hardware design, or low-level performance optimisation. Its trade-offs — density over speed, simplicity over stability — still define how main memory behaves in every modern computer.
