Home

L2 Drive is an R&D company specializing in next-generation storage technology.

The worldwide demand for storage is increasing exponentially due to many factors — the ongoing explosion of user-generated media, an increased reliance on big data and analytics, the migration of business services to the cloud, the rise of data-driven artificial intelligence, and the rapid proliferation of embedded devices (Internet-of-Things), to name a few key areas.

Yet existing storage technologies (HDDs and SSDs) have not increased capacity fast enough to meet this growing demand.  This creates a unique opportunity for new ideas and innovations.  L2 Drive rises to that challenge with its portfolio of solutions, best exemplified by its flagship technology, the 3D-actuated vacuum drive (3DHD).

Storage Challenges

DATA GROWING EXPONENTIALLY

We truly live in the digital age.  Like never before, every aspect of our lives involves data storage.  In 2017, Seagate commissioned IDC to research the global datasphere.  In their ensuing report, IDC identified five key areas that will govern data growth over the next decade:  life-critical data, embedded systems (the Internet-of-Things), mobile and real-time data, artificial intelligence, and security.  They conclude the amount of data created in 2025 will reach a staggering 163ZB (zettabytes, or trillions of gigabytes), which is ten times more than in 2016: 


Figure 1:  Annual Size of the Global Data Sphere (from Data Age 2025: The Evolution of Data to Life-Critical

To give one an idea of the enormity of this data, storing all 163ZB would require 16 billion 12TB HDDs.  Thankfully, even though that much data will be created, not all will need to be retained.  Much of it will be used only once or twice, and then replaced by smaller metadata.  Even so, IDC projects that 19ZB of storage capacity must ship through 2025 to meet the demand.  And roughly 60% of this (or >11ZB) will still use HDDs:


Figure 2:  Enterprise Storage By Type (from Data Age 2025: The Evolution of Data to Life-Critical)

This incredible demand for data storage really cannot be overstated.  In his recent article, Luke Dormehl of Digital Trends explains,

“The reason for this is the unimaginable pace at which we currently produce data. Each day, around 2.5 quintillion bytes of data is created, courtesy of the 3.7 billion humans who now use the internet. In the last two years alone, a mind-boggling 90 percent of the world’s data has been created. With a growing number of smart devices connected to the Internet of Things, that figure is set to increase significantly.”

HDDs Are Here To Stay

Some analysts in the storage industry claim that SSDs will soon replace HDDs.  While this might be true for applications that require extremely high number of operations-per-second, like transaction processing databases, it will likely never be true for mass data storage applications.

First, there is currently not enough flash production capacity to store all the world’s data; in fact only about 10% of the world’s data is currently stored on SSDs.  And at the rate that the datasphere is growing, it is unlikely that flash production will ever meet the entire world’s demand.  In fact Samsung, the world’s number one manufacturer of flash memory, is predicting that today’s flash production capacity of ~80 exabytes (or billions of gigabytes), will only reach 253 exabytes (0.253 ZB) by the year 2020. That falls far short of the 3000 EB needed by then.  (See Samsung: NAND flash industry will triple output to 253EB by 2020.)

And secondly, HDDs still boast a 10x cost advantage over SSDs per-gigabyte, and this savings should continue, especially if HDD technology can continue to evolve through innovation.

Existing Technology Has Plateaued

Unfortunately current HDD technology is not evolving quickly enough to meet the exponentially growing demand.   First, Perpendicular Magnetic Recording (PMR) technology has reached its limits.  The biggest recent gains were achieved by increasing the number of platters.  Newer technologies that use energy assist (e.g., HAMR and MAMR) continue to be delayed with no clear timeline in sight.  As a result, drive capacities have not grown at a significant rate, and certainly not at a pace to match the rising demand.

Air-Bearing Prevents Evolution

At L2 Drive, we believe the main culprit is the continued reliance on the air-bearing.  Invented by Bill Goddard at IBM in 1956 (he called it an air-head), it has been central to HDD design for the past 60 years.  It does solve a critical issue for drive operation, namely getting the read/write head to follow the disk surface closely enough so that data can be written and read reliably.  It does so passively, by balancing the aerodynamic lift against the down-force, resulting in a dynamic but predictable fly-height.  However, even though the head flies above the disk surface most of the time, it does touch the surface periodically.  And each time it touches, it scrapes away some of the lubricant layers.  That material then collects on the read/write head or gets dispersed throughout the drive.  Eventually there is not enough lubrication remaining and the drive fails due to a head crash.  Indeed 75% of drive failures are due to this phenomenon.

Our Approach

3D-ACTUATED VACUUM DRIVE (3DHD)

In order to permit HDDs to continue their evolution, the existing air-bearing needs to be replaced.  To this end, we have designed, and successfully prototyped, a 3D actuator operating in a vacuum.  The head-media-spacing (HMS) is monitored at 100 kilohertz and adjusted at 15 kilohertz using a piezo-crystal mechanism (controlled by firmware in the drive).  Furthermore, the read/write head is carried rigidly on this mechanism instead of by a flexible load beam.

3DHD Benefits


Figure 3:  Power Consumption in Air vs. Helium vs. Vacuum

The benefits to this design are numerous.  First, in a vacuum, disk flutter is minimized and air-turbulence effects are eliminated.  With no atmosphere swirling in the enclosure at ~100 mph, the drive consumes far less power.  In our tests, we have measured 50% less idle power consumption in a vacuum.

Secondly, since the head is actively controlled in three dimensions, it will never touch the disk surface.   As a result, all media/head coatings can be removed (or greatly reduced).  That means no residue will build up on the head over time, and no head crashes will occur.  That greatly improves mean-time-between-failure (MTBF).

Current HMS (with lubricant and overcoat layers):

Proposed HMS (without lubricant or overcoat layers):

Figure 4:  HMS Reduction After Lubricant/Overcoat Removal

Thirdly, as seen in Figure 4, removing the lubricant and overcoat layers permits the HMS to be reduced, allowing for increased track density and total capacity.  (The reduced disk flutter and turbulence also improve track following, allowing greater track density.)  By our calculations, existing HDDs using PMR would see an increase of ~40% in their capacity.

Finally, 3DHD actually allows the promise of both HAMR and MAMR to be realized.  Together they introduce the possibility of 4 TB per-square-inch storage density, or >50 TB drive capacity!

About Us

Karim Kaddeche

Cofounder, CEO, and Principal Inventor at L2 Drive.
Serial Entrepreneur.

John Wang

Cofounder and COO at L2 Drive.
Serial Entrepreneur.

Peter Goglia, Ph.D.

CTO at L2 Drive.
Previously: VP of R&D at Western Digital, CTO of 
Xyratex, Director of R&D at Seagate.

Steven G. Campbell

Advisor/Investor at L2 Drive.
Previously: CTO/Senior VP at Western Digital, HGST, CEO of
Magnecomp.

Contact

If you would like to contact us, please use the form below:

Or you can reach us at:

L2 Drive, Inc.
200 Spectrum Drive, Suite 300
Irvine, CA  92618

info@l2drive.com