Why your computer's memory is constantly forgetting and rewriting itself to stay alive
In short: Dynamic random-access memory is the backbone of modern computing, yet it is inherently unstable. Each bit of data is stored as a tiny electric charge that naturally leaks away. To prevent data loss, the system must perform a constant, invisible cycle of refreshing every cell, or the memory simply vanishes.
At the heart of every DRAM chip lies a microscopic architecture invented by Robert Dennard in 1966. Unlike static RAM, which uses complex transistor arrays to hold data, Dennard’s design relies on a single transistor and a capacitor. This structural simplicity allows for massive data density and lower costs, making it the standard for main memory in everything from supercomputers to portable devices. However, this efficiency comes with a trade-off: the capacitor’s charge is fleeting. Without an external refresh circuit to periodically read and rewrite the data, the information would be lost almost instantly.
The process of reading and writing is a delicate dance of electrical signals. When a row is accessed, sense amplifiers detect the tiny voltage difference from the capacitors and amplify it to a full logic level. Crucially, this read operation also serves as a refresh; as the sense amplifiers output the data, they simultaneously push charge back into the capacitors. Manufacturers typically follow JEDEC standards, which mandate that every row must be refreshed at least every 64 milliseconds to maintain integrity. This requirement demands precise timing and complex memory controllers to manage the constant interplay between data access and maintenance.
The history of DRAM is marked by rapid innovation and intense market consolidation. After the Intel 1103 became the first commercially available DRAM in 1970, the industry saw a shift from US dominance to Japanese leadership in the 1980s, followed by further shifts toward Korean manufacturers. Today, the market is highly concentrated, with only three primary global suppliers—Micron Technology, SK Hynix, and Samsung Electronics—exerting significant control over capacity. This concentration, combined with the surge in demand for AI-focused high-bandwidth memory (HBM), has led to significant price volatility, as HBM production often competes for the same wafer capacity used for general-purpose memory.
While DRAM is classified as volatile memory, it is not entirely instantaneous in its data loss. Under specific conditions, some information can be recovered even after power is removed or refreshing stops for several minutes. Despite this, the fundamental reliance on dynamic logic remains the defining constraint of the technology. As performance demands have grown, engineers have moved from asynchronous designs to synchronous (SDRAM) and double data rate (DDR) architectures, focusing on bandwidth and pipelining to overcome the inherent physical limits of the capacitor-based cell.
Opens the Intellect+ feed with this card first. Swipe up for the next one.
Sources
- Summary by Intellect · source: Dynamic random-access memory
To cite this card, link to https://intellectplus.co.uk/app/technology-dynamic-random-access-memory.
How we source, check and correct cards: Editorial standards.
General information, not professional advice. Spotted an error? Flag the card or see corrections and complaints.
Intellect Plus is a visual educational discovery platform for exploring fascinating facts and ideas. Brain Food, Not Brain Rot. What is Intellect Plus?
Related
More in Computing and information technology · More in Technology · Latest · A–Z · All topics