Seagate is bigging up HAMR (Heat-Assisted Magnetic Recording) as the replacement to take us beyond the disk capacity limits of current PMR technology.
PMR (Perpendicular Magnetic Recording) involves the magnetisation of tiny bits of magnetic material, made up of grains, oriented vertically in the recording medium. As the size of these needle-like structures decreases so too does their ability to resist magnetism changes from temperature fluctuations and neighbouring PMR "needles".
Two ways have been suggested to get around this. One is bit-patterned media (BPM) recording, which involves putting an insulating doughnut-like ring around bits laid out in a precise and very hard to produce pattern. The other is to use a recording medium in which the tiny bits have to be heated in order to have their magnetic direction changed - this is called HAMR.
Both HAMR and BPM use will involve substantial and costly development and changes to production machinery. Shingle writing (using partially overlapping tracks to increase track density) is being proposed by suppliers such as Hitachi GST as a kind of temporary fix to extend the life of PMR by increasing capacity by 10-50 per cent.
As reported in ConceivablyTech, Seagate's recording media operations VP, Mark Re, thinks PMR can get areal density to 1Tbit/in2, and we'll reach that in the 2013-2015 period.
Some current drives, such as Toshiba's 750GB, 5400rpm, 2.5-inch MK7559GSXP and Seagate's 750GB, 7200rpm Momentus, are at 541Gbit/in2 and a jump to generally available 600Gbit/in2 class drives is surely coming soon. We appear to have two or three more PMR generations ahead of us before the 1Tbit/in2 limit is reached.
Re thinks HAMR could take areal density up to 50Tbit/in2. A HAMR/BPM combination could then conceivably take areal density to even higher levels. He says that HAMR drives will use an iron-platinum alloy and standard read/write heads, not the optical technology that Seagate acquired with Quinta in 1998. Seagate has made prototype HAMR drives but these are not capable of being productised yet.
It looks as if Seagate is currently betting its disk drive farm on HAMR and not BPM. It may bring in shingle writing as an interim PMR boost but, because of its write speed limitations, this may not apply to enterprise drives. Adding more platters might be the route chosen to increase enterprise drive capacity until HAMR becomes a viable production method.
Bootnote: Seagate trumpeted its HAMR intentions back in August, 2002. It said then HAMR could achieve 50Tbit/in2. ®