Despite the common wisdom, however, some suggest that this single-minded pursuit may mean the storage industry is ignoring some pressing issues. And for some market segments, especially in the enterprise, more capacity may not necessarily be better or even necessary.
At the recent Diskcon conference of IDEMA (the International Disk Drive Equipment and Materials Association), in Santa Clara, Calif., almost every presentation began with recollections of the ancient days of hard drive manufacturing when the memory capacity that now fits in an iPod used to fill a bunker.
Often, the speakers continued the history lesson with a comparison of hard drive advances to those of some other invention, usually automotive.
For example, in his "kickoff" speech, Seagate Technology Chief Technology Officer Mark Kryder said that if a 1956-vintage standard car had undergone the same rate of "progress" as a hard disk, "We ought to be squeezing 146,800 people into that automobile today; the price should have dropped to $15; and have a top speed of almost 1 million miles per hour."
Following the wave of such stories (including his own), Dan Frost of the San Francisco Chronicle blogged a clever response from a reader, which then made its way around a number of storage lists. Heres bit of the post:
"If my car was like my hard drive, I would need to keep an exact copy of everything that I carry in the car because sooner or later the car is going to lock itself, and I will never get into it again. If I decide to go to the trouble of getting into the car, I will have to take it to a specialized mechanic who will probably charge as much as the car cost, with absolutely no guarantee of salvaging anything," the reported author, Dave Hector, observed.
His final shot was: "You get the idea. I love my car and I trust it. I love my computer, but I dont ever, ever, ever trust it."
While this comment is certainly entertaining, it is also educational. It points to a fundamental disconnect between the makers of storage and computing systems and their customers.
In the storage industry, almost all R&D is focused on raising the level of areal density in a drive, or the number of bits of data that can be packed on a disk platter. This is the way that we find ever higher-capacity hard drives in our notebooks and desktops, or smaller drives in portable music players that have the same capacity as larger drives from a few years ago. All makers of storage, including hard disk drives, flash and optical, are driving towards higher capacities.
At Diskcon, all presentations contained a mention of the continuing progress toward higher densities. There were many graphs showing how this and that technology would let a company keep up with the storage growth curve. Almost all questions raised were about the rate of increasing capacity, its sustainability and what technology will be employed to improve capacity. And whether a profit can be made in this manufacturing cycle.
But customers might have a different view of the priorities for progress in the storage arena and what benchmarks would constitute progress.
Items on a customers list might include: What have the manufacturers done to improve security on disk mechanisms? What can disk makers do to increase performance, especially seek performance, which hasnt kept up with capacity in any way? And then theres the subject of reliability.
The response from the storage execs appears to be mostly finger-pointing to different parts of the system. Or worse, just a shrug. The answers must come from someone else, they seem to say, since they are too busy working on the capacity problem.
Sure, hard disks are more reliable than they were a dozen years ago, but how much have they improved recently? Drives sure seem to fail at the worst times.
Some presentations at the conference suggested that storage makers (or their OEMs) may not fully understand that consumers may demand more robust storage than business users and may be more sensitive to failure. (The hard disk in my DVR is starting to fail, even though its spec sheet and its low annualized fail rate would suggest that the drive should work for years more. My wife is having some trouble understanding how statistics lose their meaning when faced with the performance of an individual drive.)