In one of my previous blog posts, Five Megabytes for $50,000? A Hard Drive Blast From the Past, I wrote about the progress we’ve made with storage technology over the past 50+ years. It’s quite amazing considering how far we’ve come and it was certainly fun looking back.
But perhaps what’s more fun is thinking about what’s to come in our storage future. And with storage consumption showing no sign of slowdown, whether we’re talking storage in traditional data centers or the numerous cloud storage services hosting our data, it’s clear that a strong need remains.
First, let’s talk terminology. Overall hard disk drive (HDD) capacity is increased by technology advances made which are focused on increasing areal density. Areal density, when referencing HDDs, is essentially describing how much information can be stored in a given area. The higher the areal density number, the more information we can store. With HDDs, we describe areal density as the number of gigabits (Gb) that can be stored within every square-inch on a disk platter (Gb/in2). A Seagate Cheetah 15K.7 model which stores 600GB total has an areal density of 225 Gb/in2. A Seagate Constellation ES 2TB drive has an areal density of 347 Gb/in2. Both drives use four platters to achieve their total capacities.
One continuing engineering challenge is in finding ways to increase areal density while at the same time keeping the information stable (readable/writable). To increase areal density, engineers essentially need to work with ever-smaller bits which also need to be more tightly packed together. To keep the physics talk to a minimum here, let’s just say that this challenge in increasing areal density is what is fueling work on future storage technologies.
Even with these challenges, engineers have managed ways to increase areal densities on average of about 40% annually. And the limits of today’s technology using perpendicular recording technology are thought to be between 1 to 1.5 Tb/in2. So within a few years, we’ll need to transition to another technology.
There are numerous additional technologies out there, but for the sake of this blog, I’m going to focus on the technologies that will assist magnetic storage going forward. Even 50+ years into this, there’s no better storage technology that can match HDDs for cost. Many other technologies at the bleeding-edge look promising, but they are not without their own challenges – and costs can be astronomical in some cases! We’ll talk about some of these in another blog.
Heat Assisted Magnetic Recording (HAMR)
Between reads and writes of data on a disk, the greater challenge when working with smaller/tightly-patterned bits is writing the data. HAMR is a technology that was developed to address this specifically and is what Seagate expects to implement next when perpendicular recording limitations are reached. HAMR works by adding a laser to the write head assembly which heats the media at the exact spot where data needs to be written. The heat enables the media to become easier to write to at higher densities; the writes are still accomplished magnetically. Upon completion of the write, the laser turns off, the media cools, and the media and bit at that spot will remain stable.
So what are the current thoughts as to HAMR’s areal density limits? Seagate engineers estimate that HAMR may take us out to 10 Tb/in2 or more.
Bit Patterned Media (BPM)
The concept of BPM is to use lithography (similar how computer chips are made) to create evenly-sized and distributed individual bits on the media. With the bits being more uniform in size and more evenly-distributed , the areal density can be increased with bits made even smaller and more tightly packed together. With BPM, the data can be read or written to with further precision, avoiding interference with neighboring bits.
We’ve talked about the media, but with each of these technologies that use magnetism, the heads themselves create the electronic magnetic field which is used to flip a bit to one of two directions (think of this as a digital “0” or “1” in code) to create information. So imagine further that when the bits are made smaller, the magnetic head must also change the intensity and size of its magnetic field that it creates. If too large or strong of a magnetic field, additional bits could be flipped.
Both HAMR and BPM have been longtime projects at Seagate. The company has announced various milestones over the years and both technologies will take our storage needs out to the foreseeable future into the next decade.
Seagate also recently formed the Advanced Storage Technology Consortium (ASTC) with IDEMA (International Disk Drive Equipment and Materials Association) and other founding members including: Hitachi GST, Marvell, WD, and Xyratex. The collaboration was formed to address future data storage technology challenges and work together to combine resources to address them. The goal is through the expansion of fundamental research, ASTC will collectively shorten the time from innovation to productization, while maintaining a competitive marketplace.
The need for continued storage is clear, and the pathway to advancing the technology also continues to look strong ahead. While HAMR and BPM continue development for eventual release a few years away, further down the road, we expect that combining of both methods will be occur. It is here where technologists have estimated we may ultimately reach densities of 50 Tb/in2 or greater.
To put that number in perspective, 50 Tb/in2 would provide enough storage to hold the entire contents of the U.S. Library of Congress on a single disk the size of a 30mm diameter coin. Think we may not ever need that much storage? Time will certainly tell, but when we think to how far we’ve advanced with our storage needs over the past 10 years (nevermind the entire 50+ year history of HDDs), it wouldn’t surprise this blogger if we found ways to use this capacity in our storage future.