WavebreakMediaMicro - Fotolia
As the volume of data that enterprises need to store continues to skyrocket, getting a grip on management and its impact on cloud and mobile applications has become a priority, says Mark Lewis, CEO of Formation Data Systems. In a recent wide-ranging exclusive interview with SearchCloudApplications, Lewis discussed the growing importance of software-defined storage technology trends and the impact on cloud application development.
Storage can loosely be divided into two categories: cloud-based and on-premises. According to a newly published report offered through MarketsandMarkets, the cloud storage market size is forecast to grow from $23.8 billion in 2016 to $75 billion by 2021, at a compound annual growth rate of 25.8% over that period. Despite the continually increasing need for enterprise mobility and instantaneous scalability acting as key drivers behind the growth of cloud storage, on-premises is not going away any time soon, Lewis says. A longtime storage veteran executive who previously served as the CTO at EMC, Lewis founded Formation, a Fremont, Calif., company that believes uncoupling storage hardware from storage software is not only economical, but also easier to manage while remaining compatible with legacy applications.
Why isn't all storage in the cloud?
Mark Lewis: Cloud storage is something of a misnomer in the sense that most storage still is on-premises. A lot of cloud storage other than Amazon is entities like Facebook, Google or LinkedIn. They're really just companies with their own storage. You can call it cloud storage, but it's essentially just private storage.
Regardless of the label, what is the driver today behind maintaining an on-premises storage infrastructure?
Lewis: While we do see a huge market for cloud-based storage, a lot of organizations are looking at a mixture of public and private. Over the long haul there is going to be a large need for private installations of datacenters. That may be for control, or privacy, or security, or the inherent needs of where you want to run your own application.
For those unfamiliar with the concept, what is software-defined storage technology?
Lewis: We describe software-defined storage technology as eliminating hardware from the equation. You eliminate the storage array. Now, you're going to have scale-out, commodity x86-based gear and real-life software. You can buy different brands of hardware and different disk types and you can decide on ratios between all-flash, some flash or no flash. All of the hardware-related factors that would normally lock you into buying products from one vendor -- you now get to make those decisions. It's like when you buy a PC, you choose the hardware and features you want and then the software runs on that.
What is the practical impact of that on applications?
Lewis: The question that an enterprise should ask is whether it has a storage strategy that is future-proof so it can use whatever new technology comes along without having to replace infrastructure. If you look at software development in general, you can know that things will get better without caring how. Leave that part up to WD (Western Digital) and Seagate and concentrate on building the application functionality you need. That said, there are standards in how systems interact. We can treat it like Microsoft does in working with Intel: Both can evolve separately while staying together through standards.
What is the outlook for software-defined storage technology?
Lewis: What I've seen in customer uptake and pull-through in the latter part of 2016 signals a strong 2017. Formation Data grew nearly five times more than I thought we would in Q3. The shift is that businesses are now calling us instead of us going into a potential account and pushing and having to explain. When you're a small company you get really excited when people actually reach out.
What are those forces that are driving potential customers to seek out software-defined storage technology systems?
Lewis: That's a great question. First and foremost, it comes down to economics. Businesses' data environments are growing and they are economically challenged. They've got to start developing new apps, bringing out new products; they have uncontrollable data growth, and they just can't afford it any longer. There's nothing like a good economic challenge to bring folks to the table. We offer flexibility and scalability, but at the end of the day it comes down to this: Can you do a good job storing and protecting my data, and can you do it for a lot less money?
Where is the initial focus?
Lewis: The initial focus is legacy applications. If you have a brand-new application, you're probably building it in the cloud. If it's new and it's not connected to other applications, the cloud is probably the right solution. We found that building new applications on-premises is a two-step process. The first step is to modernize the infrastructure without changing any existing apps, because it has immediate economic benefit and simultaneously preps the infrastructure for handling future applications. It's safer than changing all of the apps and the infrastructure at the same time.
On Nov. 21, 2016, Amazon again cut prices on S3 and Glacier storage. S3 is now about 2 cents per GB per month, compared with 15 cents when the service launched in 2006. Are we seeing a race to zero, and how do the economics play out?
Lewis: Storage costs have fallen faster than Moore's Law. A lot of these price changes are a simple reflection of the provider's cost to store information. As technology advances you're going to be able to store more for less. But, companies need to remember that even with the cost of storage declining, you still have to deal with the rising cost of storing and analyzing much larger amounts of data. And there is no insight to that.
As costs decline storing everything becomes an easier strategy. Isn't there an inherent danger that developers, data scientists and anyone else in IT could become lazy as there is less need in choosing what to keep, what to throw out or what to filter.
Lewis: It's a fair point. If it's cheap enough to have a lot of processing power or storage capacity, then people will be lazier about trying to manage that. It's a fact of life. If you have a lot of photos and a service offers free storage you probably don't spend time to go photo by photo and decide which to keep and which to delete. But, if you're paying for that storage, you are more likely to make a management effort.
What's your prediction for 2017?
Lewis: I think we'll start to see new technology and vendors that no one knows about start to appear on the scene in the cloud environment. They won't have an impact yet, but will start to hit the radar in the cloud market. It won't be companies we know. There will be competition for Amazon, but I don't see it from vendors we know today. The cloud is a very lucrative market and there will be startups that come into it because of the potential. I think it will be people we've never heard of.
Cloud application-development best practices
Even cloud apps need development specs
What if you need to migrate away from the cloud?