Economic Realities

Just getting to 100 to 150 Gb/in. ^[2] will, most observers agree, require solutions to many economically (as well as technically) nontrivial problems. A survey of the hard disk market to date confirms his view.

The introduction of thin film read-write heads to displace ferrite heads took nearly 10 years. Six years were required to introduce MR head-based products, while the move from MR to GMR heads took between 12 and 18 months. Vendors agree that the rapidity of changes to existing magnetic disk drive technology have been closely tied in the past to the nature and extent of disruptive changes to established manufacturing processes represented by the new technology.

Going from thin film inductive heads to MR heads entailed a number of new processes, ranging from how the media is sputter coated, to the manufacturing of an entirely new head design with different servo controls and a number of other integration issues. Delays were inevitable. By contrast, the change from MR to GMR head-based drives occurred much more quickly. While the GMR head was a radically new design, most changes were to the head fabrication process itself, which facilitated the transition. GMR heads yielded about a 2 to 3 percent performance improvement, but a 100 percent growth in capacity, which accounted for the quick adoption of disk drives integrating the technology in the market and helped to further bolster the speed of product introduction.

Vendors claim that the introduction of 100 Gb/in. ^[2] magnetic disk drives will be an incremental one. There will likely be a reduction in the 130 percent per year compound growth rate in disk drive capacity improvement as the effort proceeds to bring new capacity expanding features on line ”possibly returning to 60 percent annual growth rate by 2005. However, demand for greater capacity seems likely to continue unabated.

Interestingly, together with increased demand for capacity, there is also an expectation that the cost per GB of storage will continue to decrease with new technology. This poses a conundrum for the disk manufacturers. On the one hand, to remain competitive with possible disk replacement technologies, they are compelled to reduce the number of read-write actuators in their disk drives, rather than increase them, and to pursue other cost-savings techniques. On the other hand, these expectations also insulate the disk industry from potential competitors offering fast, high-capacity replacements for magnetic drives going forward.

In the words of one insider, "The trend is definitely toward cost reduction, which will make it very difficult for nondisk technologies to get on the tracks. To succeed, they will need to show a profit quickly and be able to compete on a cost per GB of storage with magnetic disk. More likely, they will need to carve out their own niche by appealing to a specific set of application requirements."

The same words could be applied to magnetic disk itself. For a 150 Gb/in. ^[2] disk to succeed, it will need to deliver the same (or better) cost-efficiencies as current generation products. So, some method will need to be found to offset the cost of the additional technology required to create the new disk. Otherwise, the products will sell to a niche market defined by a subset of applications and their specialized storage I/O requirements.