Read: Computer Memory and Storage: A Historical Overview and Key Concepts

Source Details

Video Title:

Memory & Storage: Crash Course Computer Science #19

Channel/Author:

CrashCourse

Publication Date:

July 06, 2017

URL:

1. Volatile Memory vs. Non-Volatile Storage

A fundamental distinction in computer data retention is between volatile memory and non-volatile storage:

Volatile Memory:

"If your xbox accidently gets unplugged and turns off, any data saved in memory is lost."

This type of memory is temporary and requires continuous power to retain data. It is generally faster.

Non-Volatile Storage:

Data written to storage, "like your hard drive, will stay there until it’s over-written or deleted, even if the power goes out."

This form of data retention is persistent and does not require power to maintain data. It has historically been slower than volatile memory.

While historically distinct, advances in technology are blurring the lines between these two categories, with modern technologies offering "gigabytes of memory, reliable over long periods of time, all at low cost."

2. Early Storage: Punch Cards (Pre-1940s to Mid-20th Century)

The earliest form of computer storage was paper punch cards, which standardized into an "80 columns and 12 rows, allowing for a maximum of 960 bits of data to be stored on a single card."

Key Example: The US Military's Semi-Automatic Ground Environment (SAGE) program, operational in 1958, was "stored on 62,500 punchcards, roughly equivalent to 5 megabytes of data, that’s the size of an average smartphone photo today."

Advantages: They didn't need power, and paper was "cheap and reasonably durable."

Disadvantages: Punch cards were "slow and write-once, you can’t easily un-punch a hole," making them unsuitable for dynamic memory needs.

3. Early Memory Innovations: Delay Line Memory (1944 - Mid-1950s)

Developed by J. Presper Eckert in 1944, Delay Line Memory offered a faster and more flexible approach to computer memory.

Mechanism: It involved a tube filled with liquid (e.g., mercury) with a speaker at one end and a microphone at the other.

"When you pulse the speaker, it creates a pressure wave... This takes time to propagate to the other end of the tube, where it hits the microphone, converting it back into an electrical signal."

The presence or absence of a wave represented 1s and 0s. A circuit could loop the signal for continuous storage.

Key Example: The EDVAC computer (1949) utilized "128 Delay Lines, each capable of storing 352 bits. That’s a grand total of 45 thousands bits of memory."

Drawbacks: It was sequential or cyclic-access memory, meaning "you could only read one bit of data from a tube at any given instant" and had to wait for the desired bit to "come around in the loop." Increasing density was also challenging. Variations like magnetostrictive delay lines emerged, but the technology was "largely obsolete by the mid 1950s."

4. The Rise of Random Access Memory: Magnetic Core Memory (Mid-1950s - Early 1970s)

Magnetic core memory became the dominant Random Access Memory (RAM) technology for two decades due to its ability to access "any bit at any time."

Mechanism: It used "little magnetic donuts, called cores." Running an electrical current through a wire looped around the core would magnetize it in a certain direction (storing a 1 or 0). Reversing the current flipped the magnetization. Cores were arranged in grids with selection wires.

Key Example: MIT's Whirlwind 1 computer (1953) was a significant early user, utilizing a "32 by 32 core arrangement" and 16 boards deep, providing "roughly 16 thousand bits of storage."

Cost Evolution: Initially "roughly 1 dollar per bit," the cost "fell to around 1 cent per bit by the 1970s." However, this was still prohibitively expensive for large-scale storage (e.g., "4 hundred thousand dollars to store a photo on core memory").

Manufacturing: Notably, it was "typically woven by hand!"

5. Magnetic Storage for Mass Data: Tape, Drums, and Disks

The need for cheaper, higher-capacity storage led to the development of magnetic technologies.

Magnetic Tape (1951 - Present): Introduced with the UNIVAC computer, magnetic tape is a "long, thin and flexible strip of magnetic material, stored in reels." Data is written by magnetizing sections of the tape and read by detecting polarity.

• Capacity: The UNIVAC's tape reels could store "roughly 15 million bits – that’s almost 2 megabytes!"
• Advantages: Tape itself was "cheap and compact," and "they’re still used today for archiving data."
• Disadvantages: Like delay lines, tape is "inherently sequential," making access slow as "you have to rewind or fast-forward to get to data you want."

Magnetic Drum Memory (1950s - 1970s): A "metal cylinder – called a drum – coated in a magnetic material," rotating continuously with read/write heads positioned along its length.

• Capacity: By 1953, units could record "80,000 bits of data – that’s 10 kilobytes."
• Impact: Directly led to the development of hard disk drives.

Hard Disk Drives (1956 - Present): Similar to drums but use "disks… that are hard." Multiple disks can be stacked to increase storage surface area.

• Key Example: IBM's RAMAC 305 (1956) was the "world's first computer with a disk drive." It contained "fifty, 24-inch diameter disks, offering a total storage capacity of roughly 5 megabytes," finally able to store "a single smartphone photo!"
• Access: Involves a read/write head moving across spinning disks. The RAMAC 305 had an average "seek time" of "around 6/10ths of a second."
• Modern Advancements: Today, hard drives can hold "1 terabyte of data" (a trillion bytes, or "200,000 five megabyte photos") and cost as little as "$40 US dollars," a "huge improvement over core memory’s 1 cent per bit!" Modern seek times are "under 1/100th of a second."

Floppy Disks (Mid-1970s - Mid-1990s): A "close cousin of hard disks," using a "magnetic medium that’s, floppy." Primarily used for "portable storage" and "became near ubiquitous."

6. Optical Storage (1972 - Present)

Optical storage uses physical divots on a disk's surface to reflect light differently, decoded into 1s and 0s.

Examples: Began with "12-inch “laser disc”" in 1972, followed by the more popular Compact Disc (CD) and DVD, which "took off in the 90s."

7. Solid State Technologies (1972 - Present)

The current trend is towards solid state technologies with "no moving parts."

RAM Integrated Circuits (1972): The first became available in 1972, costing "1 cent per bit," quickly making "magnetic core memory obsolete."

Solid State Drives (SSDs): Modern non-volatile storage, replacing hard disk drives.

• Advantages: "Because they contain no moving parts, they don’t really have to seek anywhere, so SSD access times are typically under 1/1000th of a second. That’s fast!"
• Continuing Hierarchy: Despite SSD speeds, they are "still many times slower than your computer’s RAM," thus "computers today still use memory hierarchies."

8. Memory Hierarchy and Exponential Trends

Computer systems often employ a memory hierarchy to balance cost and speed: "a little bit of fast memory, which is expensive, slightly more medium-speed memory, which is less expensive, and then a lot of slowish memory, which is cheap."

Overall, memory and storage technologies have followed an "exponential trend," similar to Moore's Law, with costs "steadily fallen, to mere cents by 2000, and only fractions of a cent today" for vastly increased capacity.