In the race for higher hard drive capacity, reaching 20 terabytes is a big deal. Demand for data storage is growing sharply, and businesses need a place to park the bytes.
Finally, after several years of industry speculation, all three drive manufacturers are either shipping—or close to shipping—limited numbers of 20TB drives.
The 2020-21 pandemic will be remembered for many things. From an industry perspective, the arrival of the 20TB hard drive will be one of them.
Seagate and The Path to 20TB
These journeys take time. Seagate started talking seriously about the feasibility of reaching 20 terabytes as far back as 2013, according to CEO Dave Mosely.
Even that was something of a step back from previous expectations. In the 2000s, Seagate chief engineer Mark Kryder foresaw reaching 40TB by 2020 and downplayed the long-term capabilities of solid state drives.
For HDD, the reality was a much bumpier ride. The acceleration in areal density—the concentration of bits stored per square inch of media—slowed during the 2010s.
But progress didn’t stall. Hard drive makers turned to adding platters to each device to boost storage capacity. Increment jumps of two terabytes became the norm.
The strategy paid off. Seagate launched its first 6TB drive in 2014 and its first 18TB drive just six years later.
While this may not match the rapid acceleration foreseen in Kryder’s law, a tripling in capacity across a six-year period is hardly shabby. And while 2020 may not have been the year of 40TB, the prospect of mega-capacity devices remains on the cards. Only time will tell how quickly we arrive there.
Doubling Down on 20TB
For now, Seagate is focused on rolling out 20-terabyte drives. “Volumes are small but growing through calendar ‘21,” Mosley told an investment conference in November 2020.
For enterprise and hyperscalers, the expectation is to supply drives primarily using HAMR (heat-assisted magnetic recording) technology. HAMR drives are already provided to select partners, and underlie Lyve, Seagate’s nascent cloud storage platform. The firm claims that HAMR will enable the quick introduction of 22-24TB hard drives as needed, though no timeline is yet provided.
Seagate is also working on SMR drives for hyperscalers, which have a higher areal density but slower write speeds.
For consumers, on the other hand, Seagate plans to mass produce 20TB drives utilizing more traditional PMR technology. These consumer drives should hit the market later this year.
“We have the flexibility to offer multiple versions of 20-terabyte drives to meet customer needs, not only with HAMR technology”, Seagate CEO David Moseley told an investment conference in April. “We plan to begin shipping a few versions of 20-terabyte drives in the second half of the calendar year”
Western Digital Diversifies Its Approach
By contrast, Seagate rival Western Digital initially touted MAMR (microwave-assisted magnetic recording) technology as the optimal solution to the challenge of areal density. But the firm is now opting for a much more diverse approach.
And this diversified approach is yielding results一in 2020, WD introduced a 20TB drive before rival Seagate. The drive utilizes energy-assisted magnetic recording (EAMR) and SMR technology. EAMR is an umbrella term which includes both HAMR and MAMR technology. While WD claims that the drive relies on both its HAMR and MAMR research, it appears to be less technologically exotic than full HAMR, which probably helped it hit the market before Seagate’s drives.
| “Our new family of products is what we call energy-assisted. And it’s a result of inventions that we came up with as we’re working on what we call MAMR technology. We continue to invest in MAMR and HAMR. And we will bring the right technologies to the market when we think they make commercial sense.”一WD CFO Bob Elau at a conference in June. |
Pushing the Envelope
WD’s diverse approach continued into 2021. The firm is now utilizing OptiNAND technology, augmenting disk drive controllers with a small flash drive to store metadata. This has allowed WD to produce 20TB drives which don’t rely on SMR.
“OptiNAND itself is a technology that sits on top of all of our regular products that we’re going to be making” says WD’s Siva Sivaram. How far can WD go with OptiNand? The firm hopes it will allow them to produce and ship 50TB drives sometime between 2025 and 2030一without using MAMR.
WD CEO David Goeckeler makes it clear that this flexible approach is intentional. Speaking to investors in September 2020, he doubled down on the company’s position that it’s not MAMR vs HAMR but an “all-of-the-above” approach.
Neither is it about expanding capacity at any cost. Instead, what makes most sense economically and commercially will guide the way forward.
“We incorporate the technology into the roadmap at the appropriate time,” he told the call. “[We] want to be very disciplined about how we do that, so that we can get predictability of product delivery and we can control the costs.”
Western Digital is certainly working hard to squeeze as much technology as possible into its new devices. In introducing its first 20TB drives to cloud storage providers like Dropbox, WD used a combination of energy-assist technology, helium sealing, and shingled magnetic recording.
This strategy has some clear advantages. Rather than holding out for a full MAMR drive, WD can pack devices with an assortment of performance-enhancing bells and whistles that allow it to market high-capacity drives before rivals.
Toshiba Plays Catch Up
For its part, Toshiba is using a mix of SMR, TDMR (two-dimensional magnetic recording) and MAMR technologies to support increased capacities in its hard drive road map. But unlike WD and Seagate, it has yet to release a 20TB drive.
Instead, Toshiba released an 18TB MAMR HDD this year, the MG09. This may not be a bad thing in the short term, as some estimate that 2021 will see hyperscalers primarily shift to 18TB drives, despite the availability of 20TB options. Notwithstanding its initial road map, it looks unlikely that Toshiba will produce any 20TB drives this year.
| Toshiba’s MG09 drive utilizes something called Flux-Control MAMR, or FC-MAMR. Targeted for cloud and data center usage, it consists of nine 2-TB disks in a helium-filled case. |
One way Toshiba may transition from 18TB to 20TB is to add a tenth platter to the drive. The development of thinner glass substrates (needed for heat-assisted magnetic recording) may allow more platters to fit into the 3.5 inch form factor.
Importantly, Toshiba is committed to increasing its market share for nearline hard drives, where it is playing catch up to its competitors. The firm has successfully expanded its share of the nearline HDD market from 13.3% in 2019 to 21% today, though it’s still well behind Seagate and Western Digital, with a 43% and 36% share of the nearline market respectively.
Nearline is where the bulk of HDD revenue is expected to fall in the next five years. Sales of consumer hard drives are expected to almost entirely evaporate in favor of the performance benefits of SSD.
Hard Drive Capacity: Beyond 20TB
20 terabytes may have been a long time in the making, but the hard drive manufacturers are eager to maintain the quest for mega capacity gains. Of the three main disk drive makers, Seagate has its fortunes most directly tied to HDD, and is particularly bullish about the technology’s future. CEO Mosley says the company already has the ability to build 24TB drives in the lab. And he remains confident that HAMR technology will eventually enable much larger capacity jumps than presently the case.
“It’s not about getting another 2 or 4 terabytes out. It’s about how do we make big breakthrough learnings” said Mosley at a recent conference hosted by Deutsche Bank. “We’re really focused on how do we make breakthroughs in the technology so that we can get another 10 terabytes on a drive or beyond that. I want 30 and 40. And we are making progress on that and I’m very optimistic about it.”
Mosley insists that HAMR, rather than MAMR, is the right technology to achieve significantly higher future capacities per device. “We know MAMR really well. It’s a viable technology, but it’s a small turn of the crank.”
Nonetheless, he acknowledged that the development costs of HAMR technology would take time to recoup. Significant shifts in TCO would not kick in until capacities hit the mid-20 terabytes.
“I think the other opportunity in HAMR is to start taking components out of some other capacity points, say the 16TBs or the 12TBs that we’re still shipping,” he added. “If we can remove components from inside these capacity points, we’re saving money.”
“That’s the direction we’re getting on. Once we see we can accelerate that, we’ll really step on the gas.”
Embracing Innovation
While HAMR may be the long-term play, it isn’t one single technological development that will usher in the era of massive capacities envisioned by Kryder’s law.
Instead, a range of engineering and software innovations are required to support and enhance the steep climb up Higher Capacity mountain.
Supporting technologies on the path to mass capacity include:
Refinements in recording technology:
There’s a number of approaches on the table here. For example, shingled magnetic recording (SMR) utilizes the partial overlaying of tracks to increase areal density by 25%, but with performance limitations.
Two-dimensional magnetic recording (TDMR) relies on multiple read elements to generate a stronger signal-to-noise ratio. TDMR delivers a 20% boost in areal density over PMR, but comes at increased cost.
Energy-assisted magnetic recording (EAMR) of some sort will be ultimately necessary for areal density to keep increasing, according to substrate maker Hoya. HAMR and MAMR are both varieties of EAMR. These use heat and microwaves respectively to make disks more receptive to magnetic effects, allowing one to write in much smaller regions.
Further in the future is heated-dot magnetic recording (HDMR), a proposed enhancement of HAMR which allows single bits to be stored on raised magnetic grains, rather than storing a bit over 20-30 grains on a continuous magnetic film.
Improvements in media capabilities:
In 2020, Japanese platter manufacturer Showa Denko announced the development of new HD media. Designed for HAMR, these platters pave the way for HDD capacities as high as 80TB, it claims.
Engineering enhancements:
Multi-actuator arms use two sets of read-write heads that logically divide a disk drive into two halves and perform read/write operations simultaneously to increase overall input/output.
Protocol and software development:
Both Western Digital and Seagate have embraced open source RISC-V processor designs to better support the compute demands that mass capacity storage will bring.
Moreover, development work is afoot to extend the flash interconnect protocol NVMe to HDD. If successful, this promises to extend some of the massive parallelism enjoyed in flash-native environments to hard disk drives, with their traditionally slower data transfer rates.
It’s an all-of-the-above approach to innovation that will help propel the hard drive makers forward.
Hard Drive Capacity: Exposing Market Tensions
Of course, the manufacturers have a vested interest in talking up the potential for massive capacity increases.
But the march toward greater capacities is not slowing up any time soon. Higher capacities are essential to preserving HDD’s cost advantage over SSD in the data center.
And it’s in the fault lines between HDD and SSD that we see the industry battle play out. Seagate, heavily dependent on spinning media, believes HDD will retain its price advantage over SSD until at least the middle of the 2030s. By contrast, Western Digital insists its growth investment in flash remains strategically correct.
“We believe flash is the greatest long-term growth opportunity for Western Digital,” said CEO Goeckeler. “It’s an area where we’ve already had a tremendous foundation— with consumer cards, USB drives, client and enterprise SSDs.”
Nonetheless, he recognizes that HDD remains critical to meet exploding demands for data storage. “Storage is the fundamental underpinnings of the digital economy we live in. Continuing to drive the roadmap of hard drives is extremely important. [HDD] is the storage of the public cloud and it’s going to be for a long time to come.”
Toshiba, doubling down on its nearline production, shares this view. Hard drives remain the most commercially viable way to store data at scale “based on current price points and future price projections.” That’s according to Rainer Kaese, senior storage manager at Toshiba Electronics Europe and long-time HDD evangelist.
“The price/capacity of SSD and the lack of adequate production output means that it will not be able to absorb the data explosion that is currently on the horizon. HDD technology is much better positioned to do so,” Kaese remarked in a December 2020 op-ed.
Game On
For each of the manufacturers, the race for increased hard drive capacity is real. For as long as serious market demand exists, the trio will run hard and fast in pursuing additional terabytes. The question remains: who will most comfortably stay the course?
At Horizon Technology, we know all about data center HDD. For practical support with managing jumps in hard drive capacity now and in the future, get in touch.
For more insights on the exabyte growth driving the race for higher capacities, read:
