I originally arranged to interview V3 systems CEO Peter Bookman at VMworld in San Francisco back in October. However, we weren’t able to schedule the time until the final day, by which point we were both so tired that we never got past the “talking about cars” stage. I finally caught up with Peter last week, and this time we managed to get past the “talking about cars” stage. I tried to limit myself to just asking questions rather than sharing my opinion, and I’ve made a couple of edits for clarity and tidied up product names etc. where appropriate, so any errors in the transcript are mine.
(Simon Bramfitt: The Virtualization Practice) First of all, why VDI? What was it that made you feel the need to launch V3 Systems?
( Peter Bookman V3 Systems) Great question. First of all, I love the idea of what VDI is trying to accomplish but completely disagree with how VDI or Desktop Virtualization is going about it. I think most desktop admins can agree that desktop management and features are getting more chaotic, not less, and that firefighting issues created by the convergence of diverse application sets and devices make their jobs much more complicated. VDI, or what we at V3 Systems call Desktop Cloud Computing, can provide a compromise with a unified managed approach where desktops in the cloud can synchronize and provide features physical desktops don’t. When this synchronization is done correctly, all the features of each device can be utilized to their fullest. This includes the ability to go “offline” when cloud connectivity is compromised and then seamlessly synchronize back to the cloud when connectivity is restored.
VDI has had a rocky start—we saw massively hyped expectations from Gartner at a time when such optimism was really unfounded, followed by rapidly declining expectations as reality took hold. At the same time, much of the reporting of VDI confused it with desktop virtualization as a whole. Do you think we are past this now, and that the industry as a whole has a more realistic set of expectations about what VDI is, and how best to take advantage of it?
What is in a name? I think a lot is in a name and VDI has been very controversial. I think we can all agree that the demand for such solutions is very real, but offerings to date have been very misleading in what they actually can do, not to mention the complexity associated with managing more things and more complicated solutions. I think we will not get past this until we simplify everything. I look at the appliance model not as a chance to rebrand servers, but as an opportunity to provide purpose-built appliances for flexible workspaces. In other words, being able to first ask the question, “What kind of devices and desktops do you use now and what is the best way to move them into the cloud?” Then using a preconfigured and optimized V3 appliance for that solution means easier management, faster proof of concepts, and as many of our customers have attested to, simply not being able to go back to the “old way.”
Some people have suggested that VDI is a lame-duck technology, and that there are few good reasons to adopt it. How do you feel about this?
I’m prone to say VDI has a lot of lame-duck technologies. One example of this is IOPS de-duping. When a solution like V3’s Desktop Cloud Computing is done right and takes advantage of local storage correctly, it becomes clear that the bottleneck never was the storage. The real bottleneck is always the CPU, just like a physical desktop, so there is no need for such technology. IOPS de-duping is one of many band-aids incumbents have used in an attempt to resolve critical issues with the archaic architectures they have bet on so substantially. Caching is another example of confusing, unnecessary technology for VDI, local or otherwise. The result is mass confusion for IT departments, which can be overwhelmed by all variant technologies and hype. This confusion could be completely avoided with the knowledge that virtual desktops can be architected to closely mirror physical desktops plus add key new benefits like failover, on-demand acceleration for 3D rendering, and many others.
We have to accept that there have been the great many VDI projects that have failed to meet expectations. Why do you think we have had so many failures, storage aside? Is it a problem of technology, implementation, or what?
I think we cannot be too surprised with the level of complexity and number of different technologies which have been required to put together and deliver solutions, when for the IT departments VDI just does not cut it. Whether you choose politics, education, or any other barrier, they exist because the value does not outweigh the complexity of the total solution. When it does, it is a simple and easy decision. I had a VMware End User Computing expert tell me they won’t even bother trying to sell to an organization that does not have a VCP-certified person with experience, and especially storage experience, with all the new View stuff you now need to know on top of all of that, and sadly that is just to get started isn’t it? Simple is better, and the appliance model—with simplified tools for cloud providers to deliver to IT departments—delivers a set-top-box-like model with on-premise appliances with all the bells and whistles being delivered from the cloud. This solves both technology as well as implementation problems.
The industry has seen many major advances in the last couple of years, moving the cost of data center infrastructure needed to deliver VDI from well over $1000 per desk to less than $200. Do you think that we are approaching the end of the line here? Is cost likely to come down even further?
I find this drive within the industry perplexing. It seems to me that this push for reduced price has a root cause in the fact that VDI never did exist in a way that would establish a strong value chain in the marketplace from a price perspective. I think this drove both consumers and vendors to bypass the growth phase of a typical marketing bell curve and jump quickly to commoditized market behavior where one would find the server virtualization and desktop markets.
So comparing the price of a virtual desktop with a physical PC is a false dichotomy?
The approach we have taken at V3 Systems is quite different. We focus on building value that can be guaranteed and be clearly understood and measured by our customers. When customers understand a computing model exists that drastically reduces the time lost for system-wide upgrades and migrations that currently cost hundreds or thousands of man-hours in lost work productivity and revenue generation capability, the discussion moves very quickly from “cost” of the desktop to “value”. This discussion moves even more quickly when the idea of a premise-based “hybrid” cloud offering is available that expands computing abilities for end-user flexible work-styles. When the discussion moves to the point where we talk about the value we bring with our support for multiple stacks and multiple protocols that enable them to deliver multi-purpose desktops to their consumers from the cloud, we are no longer competing on price. I expect in a few years we will approach consolidation, but not until we have gone through both the early adopter and growth phases for desktop cloud computing.
It’s more than just a question of containing cost of desktop delivery…
It’s only about the cost of delivery after the value of features have been defined and understood. But what is the value of the reduced patching downtime for desktops that has been eliminated by Desktop Cloud Computing? What is the value of not losing work because of a stolen or lost physical device? Once we can assume these types of benefits are part of a total solution while still being able to take advantage of endpoint features, I think cost will then come back into the picture and make a lot more sense. Until then, we are competing in a mythical technology war where we are trying to sell desktop technology to server guys who don’t want to be desktop guys anyways.
At V3 Systems, your approach is to offer standardized building blocks of compute and storage to deliver virtual desktops. How successful has that been?
Interesting, if by standardized building blocks of compute and storage, you mean offering turn-key “what kind of desktop are you trying to replace?” prebuilt pools of different types that deliver personalized building blocks of compute and storage, then I think our customers have been very happy with the solution. For example, one of our customers has a total of 450 employees, and the IT department had no real experience in virtualization before they deployed multiple pools, both non-persistent and persistent of varying types, with our solution. After the deployment, they told us they could never go back to the old way of delivering and managing desktops. While at our booth at VMworld this year, they simulated (not a demo, they were testing it legitimately while IT was out of office) having a site go down and having all the desktop pools fail over (even though there were a number of different pool architectures within our solution), and it worked flawlessly. This is just one of many examples where V3’s customers attest to how our solution changes their total desktop and application strategy to provide understood value to the users that the IT department supports every day and how it just got a lot easier.
So while performance concerns/benefits are often the most frequently argued, the real benefit comes from a higher quality of service?
As we have been talking about, performance is just the beginning. I think many people claim performance, but somehow it just isn’t enough, is it? I do think it worth noting that if in fact we should be able to agree that the CPU is the desirable bottleneck, then performance shouldn’t be measured in IOPS, and a better metric for storage would be latency. However, ideally no storage metric should be measured relating to performance. A robust offering that allows customers to pay for a complete on-premise deployment in a subscription model that allows for additional features and functions will begin to open the door to a future that will overwhelm anything that VDI was supposed to deliver. Creating a world where applications can be expanded and new and powerful policies can be adopted, such as guaranteed uptime for multiple desktop types and pools, or “follow me where I go” options for offline use or global travel or relocation, are all key differentiators with V3. I believe this gets to be extremely powerful when “follow me” is mixed with context-aware applications and information so I have what I need, when I need it, taking advantage of the ideal device for my application and information, however I most desire it to be most productive.
Looking at the industry as a whole, Microsoft has been singled out for criticism with regards to its licensing policies. Can you foresee anything that would cause Microsoft to modify its position here?
Microsoft is notorious for being the ultimate force in delivering what customers (not partners without customers) demand. When this is clearly understood, they focus, develop, and bring launch later in the growth phase and then market the hell out of it and take share during the mass adoption and other side of the market curves. I think that because VDI has not shown any real value yet, MS has little interest when compared with the MSP market that is well capitalized, has understood value, and where MS doesn’t infringe on its own market when using that license. I think a new type of license will (or must) emerge as we find actual adoption across organizations where MS can play a strong role in being where they need to be. I think they will get there quickly, and I know they see it and just like everyone else, they are waiting for the market to present itself. In the mean time, a “hybrid” desktop cloud solution might be a great way to validate without having to overcome the limitations of the existing licensing that seem to be a problem whether you look at MS, VMware, or Citrix. Really, this whole area hasn’t grown with the cloud and needs to, and it will, but not until a true distributed model presents itself and is worth encroaching on its own market.
Do you think there are any other steps that can be taken to further increase VDI adoption or shorten implementation times?
I think there are a lot of steps that will shorten adoption time, but the most important step is the same as for any early technology. As cloud providers deliver niche, packaged, cloud solutions (set-top box model) it will be sooooo easy to go from pilot to mass deployment and to enjoy all the benefits that desktop clouds can provide. Offering this box model on demand, in an even faster and better model than the cable television and satellite providers, on-premise will be possible, but you can try it out in the cloud first if you prefer. An on-demand 100-desktop PoC in the cloud that can simply be moved on site at any time changes the market dynamic very quickly. This is also the channel of the future that V3 is building. It is very exciting and turns existing resellers into trusted desktop cloud providers.
How do you see the market developing from here. It has been said that VDI just a stopgap solution, something to keep things together until everything is completely cloud and SaaS-centric. Would you agree?
This is something in the industry that I think we can align on. I think this is simply semantics. The desktop is evolving. The number of unique devices is growing, and I think we are just getting started with them. The desktop as we know it will evolve to be more cloud-usable and distributed. I think Windows 8 was a step in the right direction. I believe all desktops will have to evolve in that direction to stay relevant, especially as soon as hybrid and distributed desktop clouds are understood and adopted. A tablet simply isn’t a great content-creation device. On the other hand, niche devices for cars, kitchens, conference rooms, etc., will add even more complexity and value. It would be ideal if they were cloud-aware, context-aware, and always on-demand with “my desktop” which contains my personal layers, my company’s layers, and any other potential social-networking-based layers which might be ideal to create, learn, or educate from the growing pace of information out there.
How much life is there in the Windows desktop?
I’m not sure who gets the credit for Windows 8, but I think it’s going in the right direction. Much like Apple’s tablet when it first came out, it wasn’t very well understood. People thought it was a big iPhone. Now tablets are everywhere. I think Windows is a great platform, and perhaps the best platform for general office workers, and MS will get even better at it. I expect great things now that what people want is more understood then ever. V3 and other desktop cloud infrastructure companies will enable even more understanding and bring huge value in the future to the cloud regarding desktops, melding protocols and stacks from other clouds to make using best-of-breed technology as simple as checking boxes and quickly rebooting within the cloud. I think of this like seamlessly switching from one cloud-based desktop with ideal features for a particular use case to another cloud-based desktop with different features for another use case. This value will only begin to be understood as we go from Windows 8 to 9 and 10, and after Apple makes going to the cloud with its desktop and applications much simpler as well. V3 and its Desktop Cloud Orchestrator management software make this very easy, simple, and cost-effective.