For years, traditional platter hard drives have been largely stuck in terms of performance. While capacity has been increasing steadily, the real advancements in I/O have been in solid-state drives, particularly as NVM Express has been adopted as the standard interface. The inherently mechanical design of platter hard drives is the primary bottleneck–a drive head must be moved by an actuator across a platter to read and write data.
The increased capacity of traditional platter hard drives presents a problem for data centers: The drive density is leading to decreased performance in terms of IOPS (Input/Output operations per second) per TB. This has always been a pain point for traditional hard drives–the contents of an entire platter cannot be read at once, requiring an actuator to move a drive head around the platter to read or write data.
SEE: Comparison chart: NAS devices (Tech Pro Research)
Historically, 10K and 15K SAS drives were used to provide higher IOPS. SAS has no future for new enterprise deployments: The capacities are lower than SSDs, power consumption is higher, and SAS drives provide just over double the performance of traditional SATA drives, topping out around 210 IOPS. For comparison, a 1TB consumer-grade Intel 660p provides 150,000 / 220,000 IOPS, read/write; this is roughly a third of the cost of a Seagate 900GB 15K SAS drive.
SSDs cannot match HDDs in a cost-per-TB competition, however, and with the prospect of 20TB drives in 2020, the physical limitations of traditional hard drives with increased densities are forcing drive manufacturers to adopt hardware mitigations to continue providing the level of performance that enterprise applications require.
Dual-actuator drives as a solution to shrinking IOPS
Seagate’s solution to the problem is to introduce dual-actuator drives–which it brands as MACH.2–allowing drive arms to move independently of each other. According a blog post by Seagate Director of Technology Jason Feist, “With two actuators operating on a single pivot point, each actuator will control half of the drive’s arms. Half the drive’s recording heads will operate together as a unit, while the other half will operate independently as a separate unit. This enables a hard drive to double its performance while maintaining the same capacity as that of a single actuator drive.”
In a dramatic oversimplification, imagine the performance gains you would see in a RAID-0 volume, but in a single drive.
There is potential for growth in this design, as drive controllers and actuator assemblies develop around the paradigm of having multiple autonomous actuators. Feist’s blog post hints at this, noting that “In its first generation, Seagate’s Multi Actuator technology will equip hard drives with dual actuators (two actuators),” the description of “multi actuator,” and Seagate’s “MACH.2” branding appear to be setting the stage for drives with more than two actuators as densities increase over the next decade.
Giving a second life to a long-abandoned idea
This is not the first time drive manufacturers have attempted using dual actuators to improve performance. In 1994, Conner Peripherals marketed Chinook brand drives with dual actuators, though these are significantly different than the design Seagate is describing for release in the near future. (Of note, Seagate acquired Conner Peripherals in 1996.)
Although the Conner Chinook used 3.5″ platters, the external dimensions were of a 5.25″ drive, much like a CD-ROM drive. The Chinook design used two actuators placed on opposite corners of each other, providing two drive heads per platter surface. Two heads on a single platter introduced vibration issues under load, leading to higher failure rates. The Chinook drives had a poor reputation for the time, though their position as a niche product shielded them from the reputation that the IBM Deskstar 75GXP “Deathstar” suffered from.
For comparison, Seagate’s MACH.2 design conforms to the 3.5″ form factor, and uses one drive head per surface, eliminating the primary culprit for higher failure rates on the Conner Chinook drives.
Despite this, introducing more moving parts in a hard drive creates the potential for more vibration, which can lead to lowered performance, and in extreme cases, unrecoverable head crashes. Although, advances in disk drive and drive controller technology over the past 25 years may be sufficient to engineer around the complexities introduced by adding multiple actuators to a drive.
Seagate is confident that these advances are sufficient to ensure a working product. “The invention of head based micro-actuators has provided a significant step forward in vibration rejection for all HDDs,” Shashidhar Joshi, senior manager of technology strategy and product planning at Seagate, told TechRepublic. “Specifically within the context of MACH.2, we are employing techniques to mitigate cross coupling of vibrations and allow fully asynchronous operation.” Likewise, Seagate has committed to offering these drives under the same warranty terms as single actuator drives, with the same MTBF as those drives.