Workload Optimized Hyper-Convergence from the cloud…otherwise known as “give me my stuff on any device whenever I need to work on something”

An interview with John Morelli, co-founder and technology visionary for Sphere 3D

I first met John Morelli shortly before Sphere 3D acquired V3.  I had co-founded V3 to introduce the first hyperconverged appliances focused on virtual desktop workloads.  Before that I had been one of the first employees of Fusion IO a flash focused storage company for the modern data center (later taken public and sold to SanDisk in 2014).  Given my background in virtualization and storage, I was the right person to meet John.  His vision, and the amazing application containerization technology he created on the Microsoft Windows platform inspired me, and many others in my circle, and have changed our view on how to deliver software in the future; a future where there will be more devices per person than keys on their key ring.

Although I’ve spent quite a lot of time with John over the last year, I wanted to give people a chance to get to know the virtualization visionary at Sphere 3D. I decided to sit in interview fashion and get his thoughts on technologies and what he thinks gets us to ubiquitous computing, or a real personal cloud.

PB: So what inspired the idea of Glassware?

JM: Being mostly in the mobile industry for so long, and getting my start in providing virtualized training technologies for handset companies, I’ve been a mobility proponent for quite some time. It became clear to me a long time ago that there is a need for computing to be done wherever it is ideal and for that specific workload to be of the greatest benefit to the user.  I have always believed that the cloud is the best way to deliver these solutions; where the computing horsepower is limitless and costs can be controlled – keep in mind that back then we didn’t have cool names like “Cloud”, we called it the data center. So to give an example, a person should be able to use any device, whether on the factory floor, or diagnosing a patient in the ER, anywhere he or she wants and that person should also be able to use any piece of software on that device, whether it’s as simple as Angry Birds or as complicated as AutoCAD.

I also recognized long ago that none of the current virtualization technologies were designed for this use case. Everyone was focused on server virtualization, and doing a great job at it; but the end game was one of reducing costs in the data center and utilizing dormant computing resources. They were basically taking server software and running it on servers…just more of it. This was great but I wanted a greater challenge.

Nobody was designing technology that focused on moving desktop software, mobile applications and workstation or antiquated software to servers in a way that was specifically engineered for that purpose; so a few colleagues and I set out to do this.

PB:  How much of what you have done is based on server virtualization?

JM: The application containerization part of what we have done has nothing in common. Some of the clustering, traffic brokering, authentication and stuff is similar in approach only. We stripped out so much from traditional virtualization techniques when we started that we decided it would be easier to just start with a blank canvas.  It’s like the difference between bats and birds.  Bats evolved to fly too but Glassware flies in a much more efficient and elegant way.

PB: what was your first successful virtualization?

JM: We virtualized an Android phone. Not just a lookalike but the full functionality. It ran on an iPad. If you sat in your car it would use the iPad’s Bluetooth to pair with your hands free and actually make calls. We used to play android only games on our iPads just to show off. Our virtual Android 2.2 phone literally didn’t know it wasn’t real. It was built way beyond the customer specs.  That work in 2010 was the foundation for Glassware and that experience is why we chose to build technology that can leverage hardware for more than just computing.  

PB: What are your thoughts on the role of hardware in providing the best solution experience?

JM:  Containerization is powerful because it delivers applications as sessions instead of separate virtual machines.  I won’t get too technical but this means software can be delivered as it was written, from the hardware best suited for the application.    So current virtualization technologies virtualize hardware like GPUs and audio cards or have to use complicated workarounds to access that hardware.  Glassware can actually natively access GPUs and audio cards and other hardware without that workaround.  As you know, workarounds are inefficient and sometimes don’t work.  

PB: is what Glassware does common?

JM: of course not. Our approach is unique and the only technically viable means to deliver just about any application to any device. Until you can do that, virtualization of workloads (the stuff we use every day) will always have poor adoption. Right now many applications won’t virtualize on current technologies or are way too expensive to virtualize.  Would you buy a car that needs 4 wheels but only 3 work? Probably not, so why give up your physical environment and only get 60 or 70% of your applications in a virtual one?

PB: So you got it all virtualized; what do you envision the end user experience to be?

JM: Ideally the end user consumes the experience in a way which is both familiar to them as well as powerful.  This is where the name Glassware came from.  Your device is like a pane of glass that can see all of the applications running in the cloud. Once deployed, your device can be anything from high powered desktops to a smart internet of things device; the ideal experience gets built on the fly for the information which you would like to deliver or interact with.  Your application experience can be consistent across your devices (if you choose), and is no longer limited by the power of the device you are using or its operating system.   

PB: How much of a user’s device in their hand is integrated into the total solution of the Glassware platform?

JM:  In the end, we built cool protocols of our own that blow traditional RDP capabilities away; by projecting applications to your device but keeping end user device functionalities intact.  For example, you can actually choose to use the virtual application that the microphone on your device speaks to and let all the computing take place on an application running in the cloud, or you can use a native application such as Siri (on Apple) and dictate into the projected application in the cloud.  Whichever provides the best user experience ultimately should decide, not the tech limitations.  So basically the application thinks it’s running on a super high powered server but it’s actually in the chassis of your device, say an iPad.  It’s like Glassware merged the server and the iPad together into one device and that’s what the application thinks it’s running on.  It is way cool.

PB: Where do you see desktops fitting in this solution?

JM: We acquired V3 to integrate with Glassware to deliver an even more complete cloud offering.  We found that many customers still need and want a desktop.    I don’t envision the desktop going away for managing complete workspaces anytime soon; However, I do think individual applications being offered via Glassware either on a standalone basis, in groups as part of a virtual workspace, or streamed into a virtual desktop solution, is an absolute requirement if we want to see  the elimination of reliance on the physical desktop.

PB: Where does storage fit into the total solution?

JM: I heard our President, Peter Tassiopoulos say that if this were the razor and razorblade market, then the applications and the desktops are the razor and the storage is the razor blade.   We just cannot get enough data, and the amount we consume and create just keeps growing.  The better we are able to deliver and consume it, the more we interact with it, the more of it we create.  Containerizing applications and virtualizing desktops for cloud consumption greatly enhances the need for more storage and better, simpler storage offerings.  With the recent move to virtualized storage, lots of great opportunities open up for our customers.  There’s also some technical reasons why it’s good to be able to own your storage IP and control the storage architecture, but I will get into that later in a more detailed interview or blog post.

PB: Any last thoughts to share?

JM: I think we are succeeding in blurring the lines between what is local and what is in the cloud.  As end users we will no longer care where our stuff comes from, only that it gives me the performance I want in a way that is familiar to me; and whenever and wherever I want it! In the late 90s, we in the industry referred to this concept as utility computing. With containers for Windows applications plus virtual desktops and virtual storage I believe we are the most complete solutions for virtualizing desktop workloads anywhere.

PB: By the way why do you still call them applications and not apps?

JM: Peter, I’m a classic guy.  Although I work all hours of the night with these young kids in hoodies, I still like wearing my suit and tie.  Also I think calling them apps is a sign of how simplistic computing on mobile has gotten.  I want to bring back full featured applications as well as apps on mobile devices.

PB: John it’s always fun talking to you.  Will we be hearing more from you?  I hear other people are dying to understand more about Glassware, your history with it and where you think it all ends up.  I’ve heard some of your predictions and it would be pretty wild if we were involved in changing the face of technology like that.  

JM: Yeah, I’ve finally got some breathing room now.  You know, writing words that ain’t code isn’t one of my strengths, but I’ve got some things to say, and will start to say them shortly.  Thanks Peter for helping me get started with this interview.