This left me staggered. Fujitsu have been taking about a 5TB drive by 2010 and Seagate has mentioned 50TB by 2010 but 300TB? That's not just amazing, it's mind-bogglingly fantastical. What on earth is going on?
The french gaming mag report includes this statement: "Using a technology called heat-assisted magnetic recording (HAMR - excuse the small print), the company expects to be able to shove 50 TB of information into a single square inch of drive space, or around 300 TB of information on a standard 3.5" drive."
Ah. Some relief is at hand. Compare the frog text to some earlier Techworld text: "Seagate researchers have estimated capacities to reach or exceed 50 terabits per square inch using these technologies."
There has been, I suspect, a fairly basic b and B error. It isn't 50 terabytes per square inch; it's 50 terabits. 'Quelle différence' as the French might say. Indeed!
But still, notwithstanding 'les differences du bits et bytes' there is actually an amazingly large difference in HAMR areal densities between Seagate and Fujitsu.
Fujitsu thinks a 1Tb/sq in areal density is achievable with HAMR technology. Seagate is more bullish, fifty times more bullish in fact and is talking of 50Tb/sq in areal densities. Both are looking at 2010.
Okay, these are CTO-type bragging differences and 2010 is a relatively long way away. Which one is hitting the nail on the head with its HAMR? Both are projecting from their internal lab research developments and the difference suggests that Seagate is further ahead than Fujitsu in its HAMR developments.
Seagate has recently patented a HAMR-associated technology and has talked of a 7.5TB 3.5 inch drive. Even that is half as much again as Fujitsu's 5TB.
One thing, actually two things are for sure. Multi-terabyte drives should be possible by 2010 and SATA II will be far, far too slow. We'll be looking for SATA IV by then I reckon. Or maybe we'll need two read/write heads per platter as well as a bigger pipe to the server CPU.
Maybe HAMR stands for Have A Massive Rethink about disk I/O?