Communications of the ACM

these methods allow smaller regions of a disk to be magnetized, resulting in higher capacity. However, these approaches also boost costs, and result in density scaling gains of only about 15%.

Researchers are now exploring other technologies that could one day replace disks, tapes, and flash memory—or at least supplement them for specific uses. Underpinning this is the fact that up to 90% of the data generated by computers and other digital systems is never accessed again; it simply lies idle, consuming ever-growing mountains of storage media or servers.

Likewise, there is the issue of hard drive capacity. “The problem with today’s systems is that they deliver no more than one terabyte per square inch,” says Karthik V. Raman, a former IBM research scientist who now leads a research team at the Tata Institute of Fundamental Research (TIFR) in India.

The net effect is that current storage technologies—particularly disk-based servers and systems—consume massive amounts of physical space, particularly when they involve large numbers of devices and media, such as tapes or disks. Yet, even tape produces enormous volume of physical objects. It is estimated that humans produce approximately 2. 5 quintillion bytes of data each day and, overall, nearly 3 zettabytes of data exist in the digital world. All these bytes require increasingly large data centers that consume massive amounts of energy, along with other resources.

However, Raman points out, “Cre-
ating new types of storage with greater
capacity doesn’t solve the problem by
itself. There’s a need to develop bet-
ter ways to direct and redirect data for
faster processing.” Heckel and Grass,
for instance, have focused on using
DNA as a data storage mechanism.
The idea, first presented by George
Church, a molecular biologist and
geneticist at Harvard Medical School,
involves writing data to DNA material,
which could conceivably store that
data for hundreds of thousands, or
even millions, of years. Church says
the purpose of DNA storage “isn’t to
reinvent the hard drive, it’s to intro-
duce a medium that is ideal for ar-
chiving and long-term storage.”
Others are taking a different tack
for keeping data intact for long pe-
riods of time. For example, Peter G.
Kazansky, a professor at the Uni-
versity of Southampton in the U.K.,
has developed a method that uses
an ultrafast short-pulse laser to etch
data into the bulk of silica material.
“A single disk using this technol-
ogy could store 360 terabytes of data,
compared to a Blu-ray disk that can
store about 45 gigabytes,” he says.
Moreover, the data would potentially
stay on the disk for approximately 14
billion years. The project has caught
the eye of Microsoft, which is work-
ing to produce a commercially viable
version of the technology within a de-
cade (and its Project Silica is focus-
ing on ways to use the technology in
the cloud).

Tape Prevails

Remarkably, all storage devices and
use cases eventually lead back to
tape—at least, for the foreseeable fu-
ture. While tape is not as flexible or
convenient as hard drives, SSDs, and
other media, it remains cost-effective
and highly reliable. What’s more, ad-
vancements in tape continue to out-
pace other storage technologies. In
2017, IBM and Sony announced a new
magnetic tape system capable of stor-
ing 201 gigabytes of data per square
inch in a single palm-sized cartridge.
The technology has a theoretical limit
of 330 terabytes per square inch. The
world’s largest hard drives, on the
other hand, require twice the physi-
cal space, but hold only 12 terabytes
per square inch. The most advanced
SSDs hold about 60 terabytes per
square inch.

Many experts say the practical and cost advantages tape has over hard drives and other storage technologies will likely grow over the next several years. Tape won’t ever threaten hard drives and SSD for dominance, but it will remain at the center of storage— and provide a strong insurance policy for the likes of Google. Mark Lantz, manager of Advanced Tape Technologies at IBM Research Zurich, noted in an August 2018 IEEE Spectrum article that researchers continue to boost the density and capacity of tape, and the trend will continue for some time. “Tape may be one of the last information technologies to follow a Moore’s Law,” he wrote.

To be sure, tape remains viable and valuable—and the situation is not likely to change anytime soon. Concludes Grass, “Other emerging technologies will eventually change the way data is stored. But for archival data storage, tape is the technology to beat.”