Memristors: A Flash Competitor that Works Like Brain Synapses

Last week, I read a lot about HP's advances with the memristor. This is a new class of tiny switch that could eventually change some of the fundamental ways computing devices are designed, and I am very intrigued. In theory, at least, the new technology could allow for a replacement for NAND Flash memory, maybe for DRAM and hard drives, and maybe even for logic at some point. It's fascinating technology—but of course, the path from theory to commercial product is often longer and more complex that it initially appears.

Memristors, or memory resistors, were postulated initially by Leon Chua of the University of California, Berkeley, back in 1961. Essentially, the idea is that there should be a fourth device, alongside resistors, capacitors, and inductors. You could put different amounts of electrical current through the device that corresponded to different states, and the device would remember that state even after the current disappeared. In other words, it could function as non-volatile memory. Since then, the scientific community has periodically discussed memristors, but no big company seemed to take it seriously until HP announced in 2008 that it had figured out how to actually build the devices.

Last December, HP published a paper in the "Proceedings of the National Academies of Science" that detailed a new architecture with which researchers could build a three-dimensional array of memristors and address each element. Thus, it can store and read large amounts of information, according to R. Stanley Williams, senior fellow and director of HP's Information and Quantum Systems Lab.