Data Storage

Dr. Whitehead and Dr. Plumer use advanced computing to find new ways to pack more data storage into smaller spaces.

Data Storage research Compute Canada


Dr John WHITEHEAD data storage researchDr Martin PLUMER data storage research

Dr. John Whitehead & Dr. Martin Plumer

Department of Physics and Physical Oceanography, Memorial University

Research area
Drs. Whitehead and Plumer use theoretical and numerical models to study fundamental physics with very real world applications. For example, their research into the magnetic properties of thin magnetic films – at the core of many data storage devices – could lead to faster, denser and more energy-efficiency devices for memory and computation. Prior to joining Memorial in 2005, Dr. Plumer worked in the magnetic hard drive industry for seven years.

Research relevance
Consumers expect each new generation of hard drive to store more and cost less. But current technologies are reaching the physical limits of Kryder’s Law, the magnetic storage equivalent to Moore’s Law. Industry is increasingly looking to academic researchers with access to advanced research computing capabilities to come up with new methods to pack more data storage into smaller spaces.

Dr. Plumer, why did you decide to leave Seagate in Minneapolis to join Memorial University?
Plumer: It was definitely a big step, but the prospect of having all that ACEnet compute power right there was a big enticement. I was also given an opportunity to lead Memorial’s MSc program in scientific computing, the only one in Atlantic Canada. That program benefits greatly from the large-scale computing resources of ACEnet and Compute Canada.

The research you do is fundamental science but with very real-world applications. What challenges is the industry facing in the area of data storage?
Whitehead: The industry has been trying since the 1970s to shrink what you can put on a magnetic hard drive. Today, they’re looking for ideas to significantly increase or at least sustain even a modest growth in storage capacity over the next decade. This is not an easy problem.

You’ve been doing collaborative research with your Western Digital for nearly five years. This is a big California company. Why did they come to Memorial for help?
Whitehead: We were working on problems that were similar to problems the magnetic recording industry is facing. We also had the people expertise. Martin gained a lot of experience in magnetism and computation when he worked at Seagate. He also had the industry connections.

How is Western Digital benefiting from your research?
Plumer: The main things they want from us are the numerical models that can be used to study how many bits you can store on a square inch of a disk, called areal density, which is the driver of all this cheap storage. We start with very big calculations that need large-scale computing so you can model different scenarios. From there you make the model simpler for industry to use. So rather than making and testing 10 prototype devices, which can takes months and cost millions dollars, a company can do this on a computer in a matter of weeks at a fraction of the cost.

Could you have developed these models without access to ACEnet and Compute Canada infrastructure?
Plumer: No, not in the industrial time frames that they require, which is usually about six months. Western Digital would have dropped us long ago if we only had the computer power of our workstations.

Are there other benefits to using advanced research computing?
Whitehead: Security is a big issue for Western Digital. They have actually used ACEnet machines but before they did, they asked a lot of questions about security and were impressed enough to go ahead.

Are you using Compute Canada resources for other projects?
Plumer: We’re using ACENet heavily for a big project related to a magnetic material that’s used in hard drives for transducers (an electronic device that converts energy from one form to another). The project involves large simulations, working at the atomic level to study the thin films that go into hard drives. First we need to understand the fundamental physics. Once we do, we can make an engineering-friendly model that a company’s designer could use on their work station.

Whitehead: For me, having access to these computing resources has been transformative. And it’s not just the computing cycles, it’s also the support staff at the regional centres and Compute Canada. And it’s knowing that this capability will be here for at least another five years. That allows us to partner with industry knowing that we can commit to a longer-term collaboration.