Mixed Signals
Any sufficiently advanced information is indistinguishable from noise
Sunday 13 May 2007, 12:22 PM
New storage technology could replace hard disks...
or, if you prefer, "Could new storage technology replace hard disks?".
Physicists, especially those owned by computer companies and universities, are understandably keen to create new fundamental IT. Most physics, even if successful, has at best a tangential impact on people's direct experiences. A new display technology or something fantastic in storage, however, means fame within the community and - mirabile dictu - the faintest of chances of becoming known outside. But mostly, though, it's just so cool to work on stuff like that.
Unfortunately, the commercial world is no respector of cool physics. The latest report via the New Scientist from the University of Hamburg, for example, looks rather fantastic - Guido Meier has pushed magnetic domains through a tiny wire at over a hundred metres a second by nanosecond pulses of electricity. On the face of it, this is a much better idea than the way hard disks work: hard disks have mechanical bits whirring away and tiny physical tolerances that are most unsuitable to being rattled about in low-power portable devices. The Hamburg storage can theoretically be built into a single solid state device that's no more upset by the rock and roll lifestyle than any other chip.
But - ah, but - we've been here before. Solid state magnetic memory first turned up in the 60s as twistor memory, invented by Andrew Bobeck: magnetic tape wrapped in a helix around wires. That was commercialised and found a home in missiles and telephone switches, but was wiped out by silicon-based RAM. Bobeck went on to develop bubble memory, which had 'bubbles' of domains on a thin film magnetic substrate. It was faster than hard disks of the time and was just as non-volatile. This got even further into commercialisation, you could buy laptops and videogames using the technology and Intel - tempted by the potential for 'universal' memory that could replace all storage - had a 1 megabit part called the 7110.
But it never worked very well, and the hard disk soon overtook it in every respect. Hard disks continue to get better, while phase change memory is within a nanometric whisper of production (although Intel was expected to demonstrate it at IDF last month, and didn't) and is the current favourite for the much-desired universal technology.
It's a harsh environment for a young pretender. As another once-young pretender said this week, good luck...
Physicists, especially those owned by computer companies and universities, are understandably keen to create new fundamental IT. Most physics, even if successful, has at best a tangential impact on people's direct experiences. A new display technology or something fantastic in storage, however, means fame within the community and - mirabile dictu - the faintest of chances of becoming known outside. But mostly, though, it's just so cool to work on stuff like that.
Unfortunately, the commercial world is no respector of cool physics. The latest report via the New Scientist from the University of Hamburg, for example, looks rather fantastic - Guido Meier has pushed magnetic domains through a tiny wire at over a hundred metres a second by nanosecond pulses of electricity. On the face of it, this is a much better idea than the way hard disks work: hard disks have mechanical bits whirring away and tiny physical tolerances that are most unsuitable to being rattled about in low-power portable devices. The Hamburg storage can theoretically be built into a single solid state device that's no more upset by the rock and roll lifestyle than any other chip.
But - ah, but - we've been here before. Solid state magnetic memory first turned up in the 60s as twistor memory, invented by Andrew Bobeck: magnetic tape wrapped in a helix around wires. That was commercialised and found a home in missiles and telephone switches, but was wiped out by silicon-based RAM. Bobeck went on to develop bubble memory, which had 'bubbles' of domains on a thin film magnetic substrate. It was faster than hard disks of the time and was just as non-volatile. This got even further into commercialisation, you could buy laptops and videogames using the technology and Intel - tempted by the potential for 'universal' memory that could replace all storage - had a 1 megabit part called the 7110.
But it never worked very well, and the hard disk soon overtook it in every respect. Hard disks continue to get better, while phase change memory is within a nanometric whisper of production (although Intel was expected to demonstrate it at IDF last month, and didn't) and is the current favourite for the much-desired universal technology.
It's a harsh environment for a young pretender. As another once-young pretender said this week, good luck...
Previous
