Magnetic Core Memory
Before semiconductor memory was developed (which is used on almost all current computer designs) there were many devices that were used to store information. Magnetic Core memory was very common in computers that were built before the 1970s. The data is stored in small magnetic donuts so it does not need power to retain the data. It is possible to turn on an early computer that has been idle for 20+ years and still have the last program and data that was used still present in the system.
Dr. An Wang was issued patent 2,708,722 in 1955 for a Pulse Transfer Controlling Device used to store and retrieve information from core memory. He sold is core memory patent to IBM for $500,000 in 1956.
Jay Forrester from MIT was issued patent 2,736,880 in 1956 for core memory.
Ken Olsen was issued patent 3,161,861 for core memory. This was the very beginning of what was to become Digitial Equipment.
Click on the image for a larger view.
This is a magnetic Core Memory plane. This device is arranged in a 20 bit X 50 bit array. It is common to have "parity" bits in addition to the data bits that are used to check the data when it is retrieved.
You can see the individual donut shaped magnetic cores in this image.
The vertical and horizontal wires that go through the magnetic cores are use to select an individual core.
The diagonal wires are used to sense the 1 or 0 that was stored in the core.
Reading the data from the core destroys the information so it needs to be rewriten after it is read.