idspopd - Fotolia
racetrack memory in data storage products could be as little as five years away and will offer capacity to current HDD levels in a smaller form factor and data access times 100x faster than flash.
That is according to Racetrack Memory research pioneer Stuart Parkin, who is an experimental physicist, winner of the Millennium Technology Prize in 2014, an IBM fellow and director of research centres at Stanford University in the US and the Max Planck Institute at Halle, Germany.
Parkin believes Racetrack Memory is likely to arrive sooner than expected because silicon-based storage media, including flash, have reached limits in their development.
“Flash is basically a capacitor that works by applying voltages to it that are switched by a transistor,” he said. “That leads to a breakdown of cells and means a limited number of reads and writes and a tremendous overhead to ensure the memory cell is not written to too many times.
“Also, it is difficult to scale flash to smaller sizes. It being a capacitor, to get greater density, you have to put more charges in a cell and then it becomes more difficult to read and write. There is a lot of overhead with 3D solutions. Boosting the number of levels doesn’t buy you as much density as you’d think.”
Parkin’s argument is that spinning disk and silicon-based storage such as flash are essentially 2D methods of storage with access to a single layer of cells.
Racetrack, by contrast, allows one transistor or access point to connect to 100 bits held in one of “a forest of nanoscopic wires”, he said.
Read more on flash alternatives
- Flash storage has taken the enterprise by storm, but its days are numbered due to an unfavourable combination of technological obstacles and manufacturing economics.
- As IBM announces a $1bn bet on flash, UK-based ‘master inventor’ looks beyond flash and its shortcomings to the next wave of solid state storage.
The technology is based on use of magnetism, in which these tiny wires are coded by rotating the magnetic moment of atoms along the wire to opposing poles. Also, data can be induced to move up and down the wires and past a read/write head.
Parkin said the materials exist and the physics is known – what is needed next are companies to begin building prototypes.
“We are now at the point that silicon-based technologies are at the end of the road and companies need to invest,” he said.
“The physics is there, the concept works, the tooling is now available. If there were investment by companies then five years is a reasonable timeframe.”