Counterpoint about the future of tape - Technology Arena Disk vs. Tape

Computer Technology Review, March, 2003 by Juan Rodriguez, Kelly Beavers

With more than 70 years of combined experience in the tape industry, we have heard countless predictions about the impending demise of tape technologies for data storage and retrieval. In fact, the first time that each of us heard these kinds of predictions was during the first months of our very first jobs in the tape field.

[Kelly Beavers began his career in the tape storage industry in June 1974, the same month IBM introduced a new product called MSS with tremendous fanfare and boldly announced that it would quickly make tape obsolete. Juan Rodriguez started his career in the tape industry in July 1963 and, immediately upon starting, saw many of his colleagues re-assigned to work on disk technology that would relegate tape to relic status alongside punch cards.]

As you can imagine, it was a bit disconcerting to hear a eulogy for the industry we had just committed our careers to, but luckily those predictions were as accurate as all of the others in the decades since.

One of the reasons these predictions have persisted is that they have been the de facto marketing message for wave after wave of non-tape storage technologies. Proclaiming in broad terms that a new technology would mean the end of tape has always been the quickest way for a new storage technology to get noticed. Those marketing slogans may be provocative, but they have always interfered with a more informed discussion of the future of tape and the future of the storage industry.

The Future of Storage

When we think about the future of storage, we see the landscape the way most IS/IT managers do:

* Non-tape technologies will not displace tape because tape has strengths that will never be matched, thanks to the physical characteristics of tape and the rapid and sustained pace of innovation in the tape industry.

* Tape and non-tape storage solutions will be complementary technologies, deployed for applications that play to their specific strengths.

* The market for both tape and non-tape storage products will continue to grow because the applications that take advantage of their specific strengths will see a growing demand in the future.

These predictions are not provocative, but they are based on a common sense assessment of the characteristics of tape and non-tape technologies. They are also corroborated by the assessments (and buying habits) of so many IS/IT managers we talk to--people who have worked with multiple technologies, know their strengths and limitations, and know what works in different applications.

The primary strength of disk technology is its ability to rapidly access data that is in high demand. When the storage applications require rapid data access times above any other quality, disk has a clear advantage over tape.

But for applications where rapid data access times are not the only priority, or where rapid data access is one of many equally important characteristics needed, tape's significant advantages in terms of true cost, density, reliability, data protection and portability will continue to make it the technology of choice.

Tape's Advantage in True Cost

One of the areas where tape has a clear advantage over other technologies is cost. That may sound odd because so many disk vendors say that disk solutions are less expensive, but their claims are based on math that has a creative flair to it.

In their most commonly cited examples, vendors compare the cost per gigabyte or terabyte of a naked IDE drive to the cost of storing that same data on tape. IDE drives are cheap, which allows the math to work out in the favor of disk, but please remember that for an IDE drive to be functional in this application, it needs to be surrounded by special packaging, extra electronics, and other equipment that is not inexpensive. When you factor in these other costs, the price of disk storage per gigabyte or terabyte skyrockets and tape has a significant cost advantage.

Although the cost of disk storage is falling, tape's cost advantage continues to grow, thanks to innovations that will drop the price of on-line tape storage (including media) to less than $1000 per terabyte by 2005. That is a fraction of the true cost of non-tape storage, and it will make tape an even more attractive choice for any application where cost is a factor.

The Rapid Pace of Innovation in Tape

The discussion of cost brings up an important point about tape that is generally overlooked when it is compared to disk: there is often a mistaken assumption that tape is a static technology that is standing still while other technologies are moving forward. Anyone who has been in the tape field as long as we have knows how much the technology has changed over time.

The pace of change in the past few years alone has been very rapid, and that pace will increase in the coming years, based on the aggressive product roadmaps released in the past few months by tape storage companies including Exabyte.

Tape's Advantage in Density

One of clearest measures of the pace of innovation in the tape industry is density. In this case, density means GBytes/cubic inch or GBytes/rack u. Tape has a distinct density advantage over non-tape technologies, and that advantage is increasing. Tape companies are packing more and more data capacity into tape drives and media, and achieving greater and greater density in tape libraries with minimal footprints. Some new mid-range tape libraries are packing 60TB into a 20u rack space. In a few years, that same 20u rack space will hold l00-200TB. That is incredible density, and it is very appealing to IS/IT managers who are struggling to fit more storage and computing power into space that they cannot expand due to physical limitations or budget restrictions.