FotolEdhar - Fotolia
Peering into the crystal ball for enterprise cloud storage offers a glimpse of price drops, mergers and acquisitions, and possibly even a DNA-flavored archive that could keep data for thousands of years.
One longtime storage consultant predicted a customer rebellion against the egress fees that the major providers charge customers to extract data out of public enterprise cloud storage. Others see multi-cloud storage becoming "more real" than ever before, fueled in part by the rising use of container-based applications.
TechTarget sought out top technologists from major cloud providers and storage vendors, as well as industry analysts to identify near- and long-term trends that IT professionals might want to keep on their radar screens for enterprise cloud storage.
Here is a sampling of their predictions for 2020 and beyond:
Marc Staimer, president, Dragon Slayer Consulting: I think you're going to see a rebellion against cloud storage egress fees. You're already seeing some enterprise companies pulling their data back from the cloud, to the point where the public cloud vendors are going to have to do something about it. People have become very aware of the egress fees. The egress fees are going to shape how people use the public cloud. They either don't store data in the cloud, or they pull it back from the cloud. Or, they only put certain things in the cloud that they know they're not going to need. Or, they're going to have applications in the cloud that use it, so they don't have the egress fees. The egress fees are shaping what people are doing in ways that a lot of people didn't expect. I think egress fees may go away, because that is affecting business for these public cloud vendors.
Bill Vass, vice president of technology, AWS: You're going to continue to see price reductions in storage, primarily driven by increased density in hardware. You're going to see performance improvements, in general. You're going to continue to see the logic being moved to the storage plane and the compute getting closer to the storage plane for performance reasons.
We expect to see other types of storage seamlessly integrated with the clouds, extending storage out to the very edge, and the ability to have low-latency storage that's replicated up to the cloud. You're going to continue to see improved end-to-end management of storage. You're going to see storage colocated on the network spine for high-performance computing. You're going to see more threading and higher I/O to each instance to the storage -- so, massive parallelism and each individual thread significantly increasing the performance.
Focus on enterprise cloud storage TCO
Sameet Agarwal, vice president of engineering, Google Cloud Storage: With cloud moving toward more mainstream adoption, there will be a much larger emphasis on end-to-end solutions and TCO. In the past, cloud was mostly used for a small number of workloads and features, but now that companies can easily lift and shift to the cloud, the more important question becomes TCO. IT teams want to be able to manage complexity and multiple environments, so they need a simple solution to make end-to-end management easier. As this philosophy becomes more prevalent, TCO is less about smaller features and more about overall management.
Matt Kixmoeller, vice president of strategy, Pure Storage: We're starting to see containers get much more adoption. Customers who were using containers last year and the year before were mostly SaaS and web customers. We're now seeing much more traditional enterprises. One of the big reasons I see customers excited about containers is it truly allows them to build in a multi-cloud way. Up until now, it felt like customers were just trying to get their first public cloud working. They were mostly on premises, and they were starting to use the public cloud. But now that cloud has evolved more, I'm seeing customers start to think about lock-in and how to architect for multi-cloud. So, multi-cloud is finally starting to become a reality. In the past, it was talk but not much actual deployment of multi-cloud.
Enterprise cloud storage M&A in 2020?
Steve McDowell, senior analyst of storage and data center technologies, Moor Insights & Strategy: Public cloud vendors will offer more feature-rich storage options from an acquisition or two. The public cloud providers today offer basic storage building blocks, while allowing traditional players like NetApp and Pure Storage to build and sell higher-level storage solutions based on those building blocks. In 2020, Amazon, Google and Microsoft will start to take control of more of the solution stack and not continue to allow others to profit on their platforms. The fast-path to making this happen will be via acquisition.
Scott Sinclair, senior analyst, Enterprise Strategy Group: The leading public cloud providers have made considerable investments in on-premises hybrid cloud products over the last couple years with AWS Outposts, Google Cloud's Anthos, and Microsoft Azure Stack. Last year, Google Cloud Platform acquired Elastifile. As public cloud providers ramp up solutions for the data center, I expect to see at least one more acquisition of an on-premises storage technology company -- possibly another in the file storage arena, given the recent high demand for this technology.
Futuristic archival storage
Vass, AWS: There will be a long-term, durable archive type of storage -- DNA storage or something like that -- in the five-year horizon. Today, most storage products use electromagnetics to store [bits]. You basically have a one and a zero on a disk, in tape, in solid state, in RAM or in 3D XPoint storage. DNA is a very dense way to store information for thousands of years, as long as it's not exposed to certain frequencies of radiation. You use the DNA molecules to create chemical storage. You could have an eyedropper with the DNA chemicals in it, and you can have a little box, if you like, with distilled water in it. You can drip the eyedropper of chemicals in the order of the ones and zeros you want to store, and they self-form into helixes. They're millions of ways redundant when they do that. You can use all your standard storage stuff -- striping and RAID -- when you store to it. Then you evaporate the water away, and you have a little box on a chip that's got that storage encoded on it. It doesn't take any energy. As long as you keep it away from radiation, it stays very stable. When you're ready to read it again, you just hydrate it. Then you pull that water out, and you sequence it with a standard sequencer. You turn that sequence back into ones and zeros, and your storage is restored. There are new technologies that can read it much faster and write it much faster than traditional DNA sequencers. We're not committing to that, but that's part of what we evaluate and research.