Those super high-capacity SSDs that have become so very popular in hyperscale datacenters are about to lose out to a new HDD technology that promises storage at extraordinary rack densities and colossal streaming bandwidth while taking advantage of today’s trend towards liquid-cooled datacenters. The SSD Guy is told that this as a technology that promises to knock SSDs out of one of their key end markets.
At least, that’s the aim of the well-known Professor Job Lowe, Jr. of the University of Psalmanazar, orphaned son of an equally famous Job Lowe, who has created a new approach to HDDs that should give high-capacity SSDs a big challenge in the near future. Prof. Lowe has created a new approach to HDD capacity increases that harnesses two volume-production technologies from very different disciplines to vault HDDs into a higher realm of capacity and bandwidth.
Dubbed SLOBR (Superb Liquid Optical Bit Recording), the technique borrows from optical lithography techniques used for DUV (Deep Ultraviolet) patterning on chips down to sub-10nm lithographies, teaming them with the optical technologies already in use to produce HAMR (Heat-Assisted Magnetic Recording) HDDs.
DUV imagers take advantage of an optical trick called “Immersion Lithography,” which squirts a liquid between the imaging lens and the wafer, to somehow use 193nm wavelength ultraviolet light to create patterns down to 37nm. Multipatterning techniques then allow this 37nm light to create features half or one quarter that size. This same technique will be coupled with HDD HAMR technology, where the size of a bit is defined by a tiny point of laser light to produce track widths as small as to 7nm and areal densities ten times as high as HDDs currently support.
“In a nutshell,” says Lowe, the technology’s inventor, “an HDD using this technology provides storage at one tenth the cost of today’s HAMR drives that streams at least three times the data bandwidth, while packing ten times as much capacity into a standard storage rack. And it uses liquid cooling to boot!”
Although these drives have not yet been brought to volume production, Job Lowe tells us that he has approached all three leading HDD makers and has met with an agreeable response. “They smiled all through my technology pitch, even to the point of giggling, which I took to be a strong sign of glee based on their anticipation of the sheer magnitude of the potential market for SLOBR drives.”
As with any new technology, certain issues still need to be resolved. The main one is the penchant for the liquid to be spun off the disk due to centrifugal force, flying all over the inside of the rack and the walls and floor of the datacenter. A video illustrates this phenomenon in action. Prof. Lowe believes that this issue can be relatively easily solved by applying sufficient pressure to the liquid, which may require a bit more engineering in the liquid cooling system for the datacenter as a whole. For The SLOBR HDD, this will mean a titanium case with 10cm-thick walls, plumbed with custom-made thick-walled piping, as is shown in the photo below from University of Psalmanazar laboratory.
The coolant will also need to be kept to the same quality levels used in semiconductor processing, which is very stringent, although Job Lowe notes that “the semiconductor industry already knows and really understands how to produce such fluids. All it must do now is to increase the annual production quantity by a few orders of magnitude.”
Prof. Lowe expects for these issues to be resolved within a year, allowing these drives to be in full volume production by April 1, 2027.