Original URL: http://www.theregister.co.uk/2010/08/19/tosh_bpm_media/
Tosh builds mega dense hard drive - but can't read it yet
We're halfway there...
Toshiba has announced a breakthrough in extending disk drive capacity - sort of. It can build a platter holding four times more data bits, but it can't read it.
The company has been developing bit-patterned media (BPM), a technique to make smaller collections of the grains used to record the magnetic charge for each bit and lay them out precisely on the drive, with an insulating ring around them to make them hold their magnetic state more reliably. Tosh has achieved, it says, a 2.5Tbit/in2 areal density, using magnetic dots 17nm in size, but has only managed to detect tracks on the prototype media, and is not yet being able to read or write individual bits. This points to a read/write head technology problem.
The read/write head has to detect and follow tracks and then read and write the bits in the tracks. There needs to be a head servo pattern for this as well as a data bit pattern. Tosh says its demonstration of track detecting and following on a BPM disk is a world first and a breakthrough. Yes, indeed, but sadly it's not enough.
The news was disclosed at the Magnetic Recording Conference in San Diego. Toshiba's highest areal density in a shipping dive is 541Gbit/in2 in its MK7559GSXP 2.5-inch hard disk drive. A quadrupling of areal density would take its capacity up from 750GB to 3TB. An equivalent jump in 3.5-inch HDD capacity would see a leap to an 8TB drive.
It suggests it could get the dot size down to 10nm and the areal density doubled to 5Tbit/in2. The company thinks that it might be able to ship a BPM HDD in 2013.
Seagate has been steadily researching heat-assisted magnetic recording (HAMR) in which data bits are made more resistant to changing their magnetic charge, again increasing storage reliability as bits get smaller, and need a flash of laser-sourced heat to be written to.
It is being anticipated that a transition to either BPM or HAMR would involve the HDD manufacturers in huge outlays of money. Making BPM media involves etching a gold master platter at the nanometre level and then imprinting platter surfaces with the etched pattern using product machinery that is still in development from suppliers such as Intevac, Molecular Imprints  and Obducat.
The entire HDD supply chain of production and test equipment and components would collectively be spending colossal sums on the transition.
Head manufacturer TDK has demonstrated a HAMR-based head  but not a BPM one, indicative perhaps of Toshiba's head-related difficulties. Tosh has been working on Nano Contact Magneto-Resistive (NCMR) head technology  but this is clearly not at the same level of development as the media.
Shingle writing and STA
A temporary boost in HDD capacity is shingle writing , or shingle magnetic recording (SMR), in which recording tracks partially overlap like shingle roof tiles, so increasing track density on a platter. This would be cheaper than a transition to BPM or HAMR and could extend the use of the current perpendicular magnetic recording (PMR) technology for a couple of years, providing some financial breathing room. Hitachi GST is interested in this and it's thought that Western Digital, so far publicly uncommitted to either BPM or HAMR, may also be looking favourably at shingled writes.
If the HDD industry collectively wants cheaper components and production and testing machinery then it needs to avoid wasting money on the wrong bet, ie on BPM if HAMR becomes the preferred choice by most HDD manufacturers. To that end the HDD supply chain trade body, IDEMA, announced its Storage Technology Alliance  a couple of days ago, tasking it to devise an agreed industry road-map.
The founding members are Hitachi GST, Seagate and Western Digital, not Toshiba.
A Toshiba spokesperson said: "We cannot and will not comment on STA at this moment as we are not involved in this working group. However, we will continue to observe developments and support IDEMA as a whole. ... Naturally, Toshiba supports open standards and will partake and adopt solutions and standards determined and confirmed by the industry or industry bodies such as IDEMA."
Will Toshiba join IDEMA's Storage Technology Alliance? "[The] Storage Technology Alliance is one of many working groups within IDEMA. Toshiba is active member and/or founder of other working groups. At this moment, we are focusing on providing our full support to the working groups we are involved."
One issue the STA may have to confront and solve, is creating a level playing field.
Level playing field
Suppose Seagate has spent $3bn on HAMR and the STA decides HAMR is the way forward. Seagate says, "Yippee" and presses ahead and launches product a year or two ahead of Hitachi GST, Samsung, Toshiba and Western Digital, giving it a significant disk capacity - and hence sales - advantage. Why should BPM-favouring HDD vendors like Toshiba agree to that?
Another problem is down at the platter media recording layer production level. If HAMR is selected then BPM-focused production machinery suppliers such as Intevac, Molecular Imprints and Obducat would find their HDD market simply vanish in favour of HAMR machine suppliers. Why should they participate in an effort that could deny them a market?
The likelihood is that whichever technology is chosen - interim SMR, BPM or HAMR - the industry has to agree a timetable and some way of equalising member company investment, through technology licensing schemes perhaps, in order to prevent perceived unfair time and cost advantages to suppliers at the top of the supply chain.
How IDEMA and STA would cope with suppliers towards the lower end of the supply chain who face market exclusion if their preferred technology is not chosen looks to be a very tricky problem indeed.
STA membership is open to all IDEMA members, and the larger the membership the more standing and influence the body will have. It will need as much of that as it can get to cope with the tensions and potential fracture lines in the HDD supply chain and resolve them. ®