Sergey Nivens - Fotolia
In December 2014, enterprise storage vendor Overland Storage merged with Canadian virtualization software vendor Sphere 3D Corp. The merger followed years of financial struggles and transformation for Overland, which had already de-emphasized its original tape storage product line to focus on its NAS SnapServer and RDX QuikStor line of removable disk backup appliances.
Sphere 3D's flagship product is Glassware containerization software for virtualizing desktop applications running in Linux, Microsoft Windows and mainframe operating systems. While Sphere 3D continues to sell Overland's disk and tape products, its focus is proprietary software containers.
SearchVirtualStorage recently sat down with former Overland and current Sphere 3D CEO Eric Kelly to discuss the newly merged company's strategy for assimilating its networked storage, tape and virtualization offerings.
Overland Storage had been around since the 1980s and has a known storage portfolio. Why did you choose Sphere 3D rather than keeping the Overland Storage corporate name?
Eric Kelly: We wanted to really transform Overland from being well-known as a tape library vendor to a new company focused on virtualization and storage, with very little tape business. But it's hard to shake the brand. The good news is Overland had a very good brand and did a great job over 30 years of branding itself. The bad news is that Overland did a very good job of branding itself. We decided to keep the product brand names that are tied to Overland, such as the NEO (XL) tape library and SnapServer, but from a corporate positioning, we wanted to emphasize Sphere 3D's vision, which is the ability to virtualize any app on any device. We've trademarked that [slogan] and even changed our stock symbol of NASDAQ to ANY.
Will Sphere 3D focus on selling Overland storage hardware or shift more into storage software?
Kelly: Distributed storage architecture is becoming more and more important. That means the device you carry in your hand is the distributed architecture. That's not the traditional view of 15 or 20 years ago, where distributed architecture meant your PC or laptop. As a company, we are looking at a total addressable market that IDC estimates will top $50 billion [through 2017]. Our underlying strategy is to own the IT needed to deliver a converged storage infrastructure.
For cloud virtualization, we have our Glassware software and virtual desktop infrastructure software from V3 Systems [acquired by Sphere 3D in March 2014]. We have unified storage with our scale-out SnapServer architecture. Next is converged infrastructure with our Desktop Cloud Orchestrator software. And with our backup appliances, we are building a framework to deliver a distributed hyper-convergence infrastructure. We've made a lot of progress making sure all our products talk to one another. A good example: Our NAS can be virtualized in the cloud and also have an appliance on-premises for integrated disaster recovery to our RDX backup appliances. Our VDI and virtualization applications work on the same hardware as our storage devices.
How much of your revenue growth is forecast to come from virtualization compared to your NAS, backup and tape storage?
Kelly: It's hard to give you a ballpark revenue percentage for virtualization, but it definitely will be the face of our company going forward. And it goes across the product line. Our NAS product has to live in a virtual world. They have to go hand in hand. Our virtualization strategy is for the entire storage stack.
What is Glassware software and what are its implications for enterprise storage?
Kelly: Glassware is containerization software. If you're familiar with Docker containers, which are for Linux-based storage, it's similar … but we refer to Glassware as a 'productivity container.' Glassware is able to deliver storage for any application in its native format on any device, including Android, iPad, even virtualizing mainframe apps to a new blade server. The three main elements are economics, performance and security.
Is Glassware the Sphere 3D equivalent of a hypervisor?
Kelly: It's a thin software layer, thinner than a traditional hypervisor. We call it a 'micro-visor.' Glassware handshakes with the necessary components, but doesn't interfere with the connection between an application and storage hardware. That's our secret sauce. It's a different architecture than VMware and Citrix use. We did benchmark testing at Dell World 2014 where we virtualized more than 7,000 concurrent applications running on a PowerEdge VRTX box. The same VRTX hardware running VMware was measured at 500 [concurrent applications]. That's a 12:1 ratio of economics with Glassware for building out cloud storage infrastructure. To give you an idea of where we're going with this, we filed a patent to put Glassware on a chipset. That would mean you no longer need [products such as] Citrix and VMware.
Are you suggesting that Sphere 3D could eventually replace hypervisors? That's a pretty lofty claim.
Kelly: I wouldn't say we will eliminate them completely. I started my career on the mainframe side, and mainframes still exist, but they aren't the dominant players. I think virtualized storage is heading in the same direction in that things are getting pushed closer to end users. It's still early, but we haven't found applications that we can't virtualize. The toughest one was to see if we could virtualize a mainframe. We're pushing the limits. People talk about the 30% of apps that just won't virtualize. We actually virtualized a DEC PDP-8, which was the oldest version we could find. We used our container technology to split the DEC operating system into a container and are serving it up on our storage device.
Given your emphasis on virtualized cloud storage, when will you roll out products to support object storage?
Kelly: Object storage is an important layer because it gives you all the capability you need to search for metadata. There is a big push across the storage industry to support object storage. We definitely have it on our roadmap and see it as a huge opportunity. Right now, we are focused on opportunities in the education space, helping school districts virtualize apps with Glassware on Chromebooks and old PCs and laptops.
QuikStor RDX backup hardware faces competition from vendors at both the high and low end of the backup appliance market. With backup hardware facing commoditization by clouds, what does the future hold for the RDX line?
Kelly: The RDX sits right in the middle of the backup appliance market, which is about a $4 billion market. We are focused on use cases where storage has to be transportable. We own the removable-backup appliance market with our RDX product line. RDX works really well in distributed architectures and also in SMBs. It can be also be an alternative solution for tape. The RDX is a 2.5-inch disk drive that is available as a single appliance or as a 2U rack that can be placed in data center to do auto backup of servers. The drives are removable with encrypted cartridges, so you can take them with you, leave them in the rack or put them in a vault. We've shipped more than 700 units of RDX with more than an exabyte of storage globally. Most people don't realize how pervasive it is.
Sphere 3D hasn't publicly commented on its plans to continue supporting the Overland LTO tape libraries. What role does tape storage play in the company's evolving portfolio?
Kelly: I look at tapes as very similar to mainframes. It has a similar lifecycle. Tape is still the best and least expensive way to go for long-term archival storage, and LTO capacity improvements mean next-generation tape will be even more reliable. People talk about putting storage in the cloud as an alternative to tape, but if you want it back 10 years from now, you better hope it's also stored on tape.
Understanding virtualization based on containers
Docker heads up the containerization movement for the cloud
Cloud providers get ready for containerization