Cybrain - Fotolia
Tech analysts take great joy in tossing around buzzy-sounding bon mots that are supposed to succinctly describe the current state of affairs in IT. But, most of those catchphrases tend to confuse me more than shed light on knotty tech issues. So, I'm not sure if we're about to step onto or off the "third platform" of computing -- in fact, I couldn't tell you what the other two platforms are.
"Web-scale" is another epithet that is threatening to take over IT parlance. It's yet another supposedly new IT/data center paradigm. I'm a little lost with that term, too. Although I figure the "scale" part is meant to describe something really, really big, because you often hear "big data" cited as one reason for companies to pump up their IT operations to "Web-scale" proportions.
Say it three times fast: "Social, mobile, cloud."
And the reasons behind all this third platform and Web-scale talk are the now all-too-familiar triumvirate -- social, mobile, cloud -- that comes tripping off the tongues of marketers and analysts alike. (When desperate, the same sloganeering gang will toss in the now classic "big data.")
But what's wrong with that picture? Social, mobile and cloud don't describe applications or data or any type of intelligence. They're all places or environments where something might happen.
The emphasis seems to be on delivery, on being able to provide access to apps and data whether the person looking for those things is in the office, on the road or feverishly crunching numbers. Or avoiding that tough stuff and being social instead.
If we believe what we hear, the IT scene is increasingly about where and how, and less and less about what or why. Still, the thing that seems to be getting short shrift is the data itself, and how that data is stored.
It's almost as if it's become irrelevant where the data resides. But that doesn't make sense. We already know that all roads lead to storage in one form or another and the trick is to find the best type and configuration of storage media to keep the traffic running smoothly. But storage can be more than the repository of all that hopefully valuable corporate IP. It can be the center of a true information processing and management system.
Outside looking in
When we take a third-platform or Web-scale view, we're looking at the data and its apps from the outside, and most of the developments in storage tech that we see today are taking that external approach to solving what was an internal problem: We know we're pushing around billions of bits every second, but we don't really know what those bits are all about.
Most of those new techs claim an advantage because they can "abstract" the storage software from the storage hardware. That in itself is kind of misleading. Software and hardware have always lived separate yet interdependent lives, and what software-defined storage -- or software-defined anything -- purports to do really isn't any different functionally. What's different is the new model discourages vendors from creating custom hardware and software which yield a proprietary architecture. That's not a negligible point because it may cut costs, maybe even substantially. But it doesn't make any fundamental changes to how data is managed -- or not managed.
Without a lot more insight into the data than we currently have, we can't just keep shoveling bits around anymore. When there are more and more bits to shovel, it gets harder and harder to know where those bits might be needed, when they're useless and what they mean.
Storage is getting smarter
So-called "data-aware" storage systems -- like those from innovative newcomers Data Gravity and Qumulo -- have the potential to shake up the storage space race. They break new ground by getting inside the data to improve storage and data management.
Some of these new approaches represent significant improvements over traditional storage systems and even over the new wave of software-defined and hyper-converged storage products. But, I think we're still falling short of being truly data-aware -- it's more like data-capacity or data-activity aware.
For storage to be managed most effectively, the data needs some intelligence of its own. Sophisticated new storage systems can watch over data and see how it's being used, who's using it and how often it's being used, but that's probably not enough intelligence to make more incisive determinations like when it can be used, when its usefulness has expired, under what circumstances it should be used and so on.
Data storage Tower of Babel
In theory, the storage system could collect that information and make it available. But, in the data processing cycle, that type of intelligence typically comes from the apps and users of those apps. Today, storage systems can't collect that data, because apps, OSes, hypervisors, storage systems and so on, speak different languages.
Since IT began, there have been countless attempts at developing a lingua franca so apps and data systems can talk, exchange data and even share information about that data. Some have been successful like XML and DICOM (Digital Imaging and Communications in Medicine) while others not so much, like SNIA's Cloud Data Management Interface. But for storage to really get smart enough to be an active partner for managing information, app makers and storage vendors -- and everyone in between -- will have to start sharing what they know in a format that they all can understand.
Until then, we'll have to occupy ourselves buying bigger and bigger disks, and building infinite clouds. Who knows -- maybe common sense is just ahead waiting for the rest of us to arrive on platform number four.
How the "third platform" affects the data backup process
Taking another look at IT's third platform