WavebreakmediaMicro - Fotolia
- Mike Matchett, Small World Big Data
Have you sensed a shift in storage these days? Maybe you've noticed a certain resignation among storage industry veterans contemplating the future of data storage technology. Or maybe when it comes time to refresh aging storage arrays, there's less differentiation among competing products or no exciting new storage technologies -- everyone has flash by now, right? -- or flaming vendor wars to get excited about. Maybe many of your important storage needs are now met using a cloud service, a relatively no-name vendor product or even open source.
For many years, it's been fun to watch the big storage vendors fight the good fight. They used to line up elbow-to-elbow in the front row at big shows like VMworld, vying for the biggest booth to show off their hottest products. This last year, it seemed storage has moved back a few rows. Market forces and trends such as software-defined and hyper-converged have changed large parts of the storage game, sure. But when the game shifted in the past, competitive storage vendors shifted with it. Maybe this is harder to do now that storage is getting embedded, integrated, converged and "cheapened" through cloud competition.
Many recent storage trends involve getting rid of storage as an IT silo, raising questions about the future of data storage technology. How can you sell Storage with a capital S if no one is buying stand-alone storage anymore? Are we coming to the end of storage as an important, first-class industry? The short answer is no.
But data? Lots of data
Accounting-focused industry reports show legacy storage-centric companies continue to suffer from thinning margins for their high-end hardware arrays. But the collective storage footprint in general is growing. With data volumes exploding from globalized applications, web-scale databases, big data analytics, online archiving and that little internet of things opportunity, all those new bits will have to go somewhere.
All this new data simply can't go into cheap and deep cold cloud storage. If data is worth having, as much business value as possible must be wrung out of it. And if it's important data, it has to be governed, protected, secured and ultimately actively managed. Even if storage gets molded into converged, hyper-converged or cloud infrastructure, much of it still holds important corporate data.
It would seem the future of data storage technology will see storage vendors continuing to compete, differentiate and even excite their customers with ever-increasingly powerful data management services. Many advanced services are layerable, however, meaning they don't tightly integrate with storage arrays or lower-level storage components. This lets third-party vendors swarm in, and while that might be good for consumers, it's tough for storage vendors basing their business on higher-margin assumptions.
Some advanced data services can benefit from tighter, low-level integration in storage, of course. But these services and the vendors that offer them may find the best opportunity to shine when deeply converged with more than just storage components. The trouble for pure storage vendors is that hyper-converged products, as they're integrated with compute, networking and other areas, aren't confined within the storage silo and take visible market share away from the storage market. So, for example, Hewlett Packard Enterprise's Simplivity hyper-converged product isn't in the HPE storage division.
Set storage to automatic
When it comes to the future of data storage technology, maybe storage management is more the issue than the changes in storage delivery. Is increased IT automation the real culprit here? Maybe traditional enterprise storage and all its expertise-demanding administration and management are destined to become just a fully automated component of more cloud-like IT architectures. If this is the case, should all of us storage folks simply close shop and go home?
Now don't get me wrong, IT automation is a good thing. It's at the heart of most new IT initiatives, including orchestration, self-service and DevOps. Increasing automation is a no-brainer; IT should always strive to automate repetitive, onerous and error-prone tasks. And vendor-validated automation enables the embedding of many storage management tasks opaquely inside larger systems, effectively hiding any storage complexities.
Therein is a ray of light for storage professionals. When an organization writes and deploys an automation script, it incurs a bit of technical debt. At some point -- or even many points -- in the future, that automation likely must be optimized, enhanced, debugged, profiled or evolved.
A brave new storage future
This might be unpopular to say, but I think valuable storage expertise is shifting from deep trivia-based administration to a style of dynamic operational management closely related to DevOps. We might call this future role OpsDev, in which the now agile infrastructure operator dynamically programs and reprograms an increasingly software-defined IT stack, including storage and data management services.
With vendor call-home support schemes becoming more dynamic and real-time with every release and machine learning algorithms included in just about every facet of IT systems management, you can argue that stand-alone storage expertise will soon become unnecessary. But no matter how smart or clever remote management becomes, an on-site professional who knows the business and its data requirements, who knows what's acceptable to the users and other specific requirements, will have ownership over corporate responsibility.
We still have a big challenge on our hands, though, when it comes to the future of data storage technology. Storage soon may be less about picking out specific arrays with new features every three to five years and more about assembling the most cost-efficient, scalable, globally distributed set of storage services on a daily basis. And we'll have to do all that while delivering interactive performance, instant access to big data sets and on-demand capacity, as well as maintaining full corporate governance, data protection and security.
Maybe the job won't even be called storage administrator in a few years, but rather something like chief data enabler. Who knows? The point is this is really just a turn of the big wheel of storage evolution. No matter what you call it, storage matters now more than ever. And tomorrow, it will matter even more.