Not sci-fi: A world where storage devices never fail

Last week, I wrote a piece on the future of storage. My original title was “The post-storage world,” though that was changed in edit and I did not see the change before the piece ran. The headline of “Good-bye — and good riddance — to spinning disk” didn’t necessarily match the content, which was explicitly discussing what I termed the post-storage world: that is, a world without the constraints of today’s storage mechanisms.

That’s the only reason I can fathom for the response I received in the comments, blogs, and emails. It seems that many people took issue with my essay and were eager to tell me that hard drives are cheaper than SSDs, SSDs wear out too, and until there’s a major change in how storage is handled, we won’t see spinning disk go away — it will at least be relegated to niche applications. Yes, I know. In fact, it was precisely what I discussed and what the piece was about. I wasn’t talking about tomorrow. I was talking about a decade from now.

[ Also on InfoWorld: Check out the year’s top hardware, software, development tools, and cloud services in our slideshow, “InfoWorld’s 2013 Technology of the Year Award winners.” | Get the latest practical data center advice and info in Matt Prigge’s Information Overload blog and InfoWorld’s Data Center newsletter. ]

A few folks seemed to understand where I was going, so perhaps I should try to make the idea clearer. Allow me to quote a commenter from last week:

The day the mechanical hard drive ends forever will change computing overnight.

I think the sea change in computing relevant to storage will happen before the mechanical hard drive is completely extinct, but his concept is right on. If we remove the threat of storage failure, remove the physical requirements of spinning disk, and simplify storage to the point where it becomes completely maintenance-free, we open the doors to a cornucopia of new technologies, products, and possibilities.

Through the years, computing has worked through a number of bottlenecks. For a while, it was processing power. For another while, it was maximum addressable RAM. Then it was local disk speed, local network, network storage, Internet connectivity speeds, then back to local storage again. Anyone who works in computers knows you can have the fastest processors and tons of RAM, but the computer will seem sluggish if it has slow disk. All of those CPU and RAM resources are for naught if you can’t feed them data from disk fast enough. It’s simply the laws of physics. Those bottlenecks continue to trade places as advancements are made here and there.

In the SSD era, we’ve only recently seen the effects of extremely fast local and centralized storage. We have laptops that boot in seconds from local SSDs. We have SSD-based storage arrays that can actually make use of the 10G and faster links that attach them to the network. We are now tasking our processors much more efficiently and gleaning more raw computational power out of our computing resources than ever.

That’s the tip of the iceberg — a stepping-stone on the path to storage that is so plentiful, reliable, and ubiquitous as to disappear from the mind of the general computing public. Worrying about hard drive failure will seem as quaint as worrying about a fold in your 5.25-inch floppy.

Think of it this way: 10 years ago, nearly all hard drives sold were SCSI or PATA disks. SATA drives were introduced in 2003. It wasn’t until 2005 that we saw the first 500GB disk, the Hitachi 7K500, coming in at around $350. The same bread will get you a 4TB Hitachi disk today, a drive that is not only five times larger but exceptionally faster. Of course, $350 today will also get you a 500GB SSD. Assuming there are no blocking problems along the way, extrapolating that out another decade brings us to at least the post-SSD era, but more likely, even further. After all, the computing world was abuzz with the notion of 120MB hard drives in 1993.

Finally, there’s the notion that if we have enough bandwidth, none of it matters. While broadband is still not nearly as fast as it could be, we have some residential broadband connections that are many times faster (asymmetrically, however) than the fastest LAN of less than 20 years ago. We’ve even made our first real strides toward residential gigabit Internet connections. If we can finally get to a place where most Internet connections are symmetric, or at least capable of reasonable upstream bandwidth, then remote storage becomes much closer and further blurs the lines between local and remote. This is true both for the consumer and for IT infrastructures.

If we consider the past to be prologue, then we should believe storage advancements over the next 10 years will be more revolutionary than the last 10. We don’t have much choice.

This story, “Not sci-fi: A world where storage devices never fail,” was originally published at InfoWorld.com. Read more of Paul Venezia’s The Deep End blog at InfoWorld.com. For the latest business technology news, follow InfoWorld.com on Twitter.