Top barriers to software-defined storage adoption
At this year’s Gartner Data Center, Infrastructure & Operations Management Conference a gruff, unshaven, slightly agitated audience member impatiently queried the session panelist, “Just what is your definition of software-defined storage (SDS) anyway?” A 15-minute discourse ensued. When looking to identify the factors that hamper progress of the SDS approach to storage, disagreement over the definition is just one. Below are what I believe are the most significant industry-wide barriers to widespread SDS adoption.
Market confusion, lack of education and lack of clarity
As what I saw at Gartner Data Center revealed, it’s still difficult for the community to agree on a single, unifying definition of what software-defined storage is. I’ll have to admit, technology vendors and marketers are the main culprits behind this confusion. When a sexy new category gains steam in the market, it’s all too easy to re-spin your solution as a part of the movement (can you say cloud?). The problem now is that many different types of products, including those using proprietary hardware, paint themselves with the SDS brush.
Traditional storage arrays run software code – so it’s SDS – right?
Hyperconverged appliances run scale-out software – so it’s SDS – right?
Storage virtualization abstracts networked storage – so it’s SDS – right?
The light bulb went on for about half of the visitors to our booth at the conference after we articulated the benefits of SDS. There were many however who expressed that they were not so unhappy with their existing solutions however – and certainly not enough that they felt compelled to rip and replace their entire storage infrastructure.
One executive stopped by our booth three times as he mulled the possibility of switching to SDS. His biggest barrier had nothing to do with speeds and feeds. Rather, the hang-up was his company’s organizational and team dynamics. By switching to SDS, his current storage team would likely lose members since some positions might become redundant or some would be unhappy with a new approach to storage. Others would have to organize and collaborate with other teams like they never had before. These factors were sufficient to dissuade him from SDS and why he felt it was OK to stick with his current approach to storage. The key challenge to SDS adoption in this case is simply organizational inertia – some enterprises will simply be unable to get out of their own way.
Another source of inertia is perceived risk. While companies are willing to work with emerging vendors on some datacenter components, like networking, there is still a good deal of concern that giving a startup the keys to your core data storage infrastructure is risky. Startups get acquired and distracted – or worse – can go out of business. Then what? I totally understand the concern and caution, but I also think of great companies like Data Domain who persevered, proved their worth and changed an important segment of the industry.
The idea that commodity equals cheap
One of the major stumbling blocks for IT departments and CIOs when it comes to evaluating SDS solutions is the perception that commodity servers are cheap – as in low-quality cheap. I recently wrote a blog post about this dilemma but it bears repeating. A key benefit of software-defined storage is, yes, that it can run on commodity hardware. The difference between these commodity devices and proprietary hardware these days comes down to cost.
When a major storage vendor sells you hardware, more-often-than-not what you’re buying is commodity components bundled with a few proprietary features (usually enabled by software!), and a sleek bezel (the “bezel tax”!) at a significant markup. Because the hardware ends up being very vendor-specific it adds a certain level of risk. If something breaks with a proprietary “box,” or if you need more capacity, it can mean delays while waiting for delivery. If something breaks with an industry-standard, commodity server, it can be easily replaced. The right SDS solution accommodates failures gracefully and assimilates newly added nodes and components, automatically. In this way, commodity hardware lowers the risk of downtime because it’s inherently easy-to-obtain and interchangeable without disruption. This dynamic is well understood and taken advantage of by web-scale companies, but is still foreign to traditional IT shops.
Flexibility, network quality and why x86 alone won’t save you
Hedvig customers span a variety of industries and have a range of needs. However, what they have in common is the need for a flexible storage solution that keeps up with the ever-changing demands of their data centers. In conversations with our customers and partners over the last few months, I’ve identified two technical issues that also prevent a company from getting the most out of SDS – and sometimes even stops them from moving forward altogether:
- It’s the network stupid: Software-defined storage obviously requires a network to connect all the pieces together. The speed and efficiency of this network is paramount for improved storage performance and it is usually the #1 thing we troubleshoot for customers. If you are looking to go SDS, be sure to remember that the quality of your network is as important as the SDS software itself.
- x86 will not save you: You’ll often hear the phrase “you can deploy it all on x86!” However, if it sounds too good to be true, it usually is. What I’m referring to is that hardware still matters. Deploying our software on a cluster of five-year-old servers with slower drives and anemic resources will give you a bad impression of SDS. Yes, you can deploy SDS on any brand of x86, not just any x86. To be happy with your results, mind the specs.
We are on a mission to smooth the path to SDS. Identifying, demystifying, and stomping out barriers – or perceived barriers – to utilizing the new technology is important to Hedvig. I’m sure there are other issues and concerns that exist for some of you with regard to SDS technology. I’d like to hear them. If you see other key barriers-to-entry for SDS, don’t hesitate to reach out and let us know.
To find out more about how Hedvig makes switching to SDS easy, tap the button below.