Milestone Group Quarterly: October 2008
Articles
Face to Face:
Chetan Venkatesh, CEO, Atlantis Computing
Milestone: Tell us about Atlantis Computing.
Venkatesh: Atlantis Computing is focused on the virtualization space. We are solving an important set of problems around what we call ‘the imaging of virtual machines.’ Virtualization is focused on the ability to emulate the physical infrastructure of a server or a desktop and run a virtual machine on top of it. Until now, nobody has dealt with the process and management of putting virtual machines together and then converging that with existing server and desktop management processes. That’s what Atlantis solves in a comprehensive way.
Milestone: Can you share with us your view on the virtualization market, on both the server and desktop sides?
Venkatesh: The first thing to understand about the virtualization market is that it’s in the very, very early stages. We’re in the first or second cycle of creating value and there are dozens more cycles to come. Even though we have seen aggressive adoption, virtualization is fundamentally a new, enabling technology and it’s going to play itself out over the next 10-15 years.
It’s a very simple idea to abstract the hardware from the software and allow a layer of mobility between the two by adding virtualization. That allows us to take what has been a very tightly bound hardware-software stack and loosen it up. When you do that, things like cloud computing suddenly become possible, with the obvious implications to both servers and desktops. Over the next five years, we’ll see traditional computing giving way to a model where the storage, the creation and the generation of computing power are achieved through virtualization technology.
Servers will be at the center of the cloud, managing services and desktops will be at the edge of the cloud allowing you to consume the services that are on top of the cloud. But there are enormous challenges in getting there. I think 50-70 percent of the technology that needs to exist for it to happen are in the process of being invented. So, there is just an enormous opportunity here for innovation, for new companies, for new ecosystems and it’s a great place to be at this point in time.
Milestone: What do you mean exactly when you say abstraction?
Venkatesh: In a traditional computer, the code runs on a chip, the physical form. And the abstraction is that there is no physical chip anymore, there is a software chip that emulates the physical entity. So, that's the abstraction, the fact that your end-application, or stack, doesn't run directly on the chip or on the hardware anymore, it runs on an emulated layer.
In the broader virtualization market, VMware and Citrix XenServer are emulating the physical system. But that’s not nearly enough. What we are seeing is that every aspect of the computer needs to be virtualized – the I/O, interconnects and the networking levels of the systems and the way they are composited into the processes.
Milestone: Microsoft has been very aggressive with the disruptive pricing for Hyper-V at $28, what effect do you think this pricing has on players such as Citrix, VMware, and others? And, how does that affect Atlantis?
Venkatesh: Well, it’s disruptive only to the extent that they are the first to lower the price. But pricing is dictated, to a large extent, by the commoditization of technology. You cannot charge more for something that is becoming an easily available commodity and base virtualization technologies are easily available, with very little differentiation between a vendor’s platform. You can go to the Net and find a competent virtualization platform from open source projects and build on top of it. That’s pretty much made all the offerings in the market the same.
When Microsoft decided to leverage their model by extending a very cheap hypervisor, it was disruptive more from a marketing, packaging and distribution standpoint, but not necessarily from a technology standpoint. This shifted vendors from focusing on the hypervisor to talking about how they are going to add value on top of the hypervisor.
That’s what Citrix and VMware have had to go out and do. Citrix has traditionally been very strong in managing the desktop. They have a great platform, very broad and deep, so they are at an advantage over there. And VMware, I think anticipated a lot of this happening, and have been making acquisitions over the past 18 months. And their challenge now is to integrate all of that and move up the value stack to offer more than virtualization management capabilities for the platforms, on a much broader scale. What’s needed, though, is an ability to virtualize the entire infrastructure for Fortune 500 organizations; being able to provide different models of consumption, such as private clouds and public clouds.
That’s good news for Atlantis because we are in that management space and we focus on providing a level of sophisticated management in a deep challenge area for the large enterprise. We have a tool chain to automate that. So, all of the commoditization talk has customers focusing on the real problem they have in their operations – the scale of virtualization; not just feature, function and speed.
Milestone: It seems natural that desktop virtualization and desktop management would converge. Is this happening?
Venkatesh: Absolutely, and that’s the premise of Atlantis Computing. When you are no longer constrained by the physicality of the environment, you no longer need one server per application and you can build a lot of servers out to provide high availability. So, everything changes from the kind of infrastructure that you put at the physical level, but nothing changes from a configuration and management perspective, because these are still computers except that they are virtual. It’s an everything changes-but nothing changes model.
This is one of the reasons why virtualization is adopted -- companies can see a move to this new model without too much risk. They don’t have to change their management practices or their tools and vendors. But, that only works for a limited period. When you start to seriously scale-up virtualization it inherently creates a new set of problems. So, while virtualization makes it really cheap and easy to turn up new servers, add capacity and serve customers, it puts an enormous amount of pressure on the management infrastructure. You were already challenged in managing 100 servers, now by adding a layer of virtualization they become the equivalent of one thousand. But your tools and methodologies are only capable of handling 100 servers.
Milestone: What about image management? Who do you see most active in that space and who do you think will win?
Venkatesh: This is what we do and that’s primarily why customers use us. All the different hypervisor platforms that virtualize their desktops and servers are a virtual machine image, and managing that image is a problem.
VMware has made announcements about working its way into the space in the 2009-10 timeframe. There are a couple of early stage startups that are talking about bringing some application virtualization technology and retooling it for image management and some approaches over there, but really it’s very early ecosystem, it’s formative. I think you will start to see a lot more focus on it in the middle, (next year in '09- '10) when companies start to talk about how to cut down the amount of management challenge they have with virtualization and start to adopt image management to do that.
Milestone: Do you see virtualization or part of it going and staying in the cloud? Can you elaborate on that?
Venkatesh: The cloud gives the ability to move workloads around by separating the application from the underlying hardware. Virtualization is stackable technology, so you can scale out, scale in, or scale wide. So, when we talk about the cloud as an aftereffect, it’s really putting the services and a management layer around virtualization and making it a self-service environment. I don’t see them as two separate items, I see them as one and the same.
Milestone: Our focus this quarter is on innovation. What do you see as the innovation imperatives for technology businesses over the next 12 months?
Venkatesh: It's not an incremental story any more, you need to have a disruptive advantage and be able to support that with a layer of risk management. The challenge for technology companies is in putting products inside a very, very complex operating environment for customer. And, I think a lot of innovation and a lot of innovative thinking needs to go in to be able to come up with insertion strategies that amplify the everything changes-but nothing changes model, so that customers see benefits early in the cycle of adopting the technology.
Milestone: You've certainly been involved in several innovations throughout your career. What is the essential trait required of the innovator today?
Venkatesh: I recently gave a talk to companies in India on this topic. I think you need to be able to absorb a lot of very different data and information and be able to organize it in a way that makes sense. You’ve got to be able to connect the dots across very distinct and separate domains, and spot the opportunities for innovation within them.
For example, the imaging technology that Atlantis has developed exists across four or five different domains. So, there is one part of it which is the virtualization of hardware, there is one part of it which is the virtualization of storage, there is another part which is the virtualization and transport of the network layer. And, it’s when you have enough data across all these different points that you can bring innovation to the market.
In addition to being able to ‘connect the dots,’ you have got to be a strong communicator. You have to be able to express an idea in different ways to different sets of people, and use feedback to refine the innovation. That’s what is unique about Silicon Valley; it’s an extraordinarily open ecosystem where people share ideas across a lot of different kinds of boundaries. They are very open about discussing innovation technical approaches, marketing approaches, and business ideas.
These intersect at the customer level. Today, the definition of customer has changed. A customer is anyone who is part of a community. It’s all about understanding what communities and what context your technology or innovation fits in and using the best community model for enabling that. For some communities, open-source makes a lot of sense. For some communities the whole idea of a network of people, a social network if you will, makes a lot of sense. Or in others, it’s the classic enterprise model of building a product and then selling it through a channel to the business. You take the product and service to the one which makes the most sense.
Milestone: What advice would you give to an entrepreneur who is going to jump into the innovation environment of today?
Venkatesh: Well, I think that companies will need to have strong proof points before they can call themselves a business, or decide to go down the path of taking ideas to customers. The environment is not very conducive anymore to thinking that you can throw some noodles on the wall and see what sticks. It’s a data-driven, milestone-driven approach where an entrepreneur sits down and logically thinks about the defendable milestones that prove an idea is good enough to build a business around.
I tend to think that financing organizes itself around solid, proven ideas and solid teams. So, finance just finds its way. There are smart guys who know how to deploy money and they will find you. They will hear about you if you are doing these three things. You don’t really need to go and focus on the finance, focus on these parts.
Milestone: What's one mistake that an entrepreneur shouldn't make during this process?
Venkatesh: I think the chief mistake is in not understanding the insertion strategy needed for their product or service. A lot of entrepreneus drink their own Kool-Aid and end up underestimating the resistance to change on the part of customers.
Think of the company that is running billions of dollars in sales. The last thing they want to do is rock the boat when things appear to be working fine. And as much as you can convince them at an academic level that there is an upside to rocking the boat, if you miss what they clearly see as the downsides (and how you are going to minimize them or give guarantees to avoid them), you’ll get nowhere. I call the ability to address this resistance an insertion strategy.
Entrepreneurs are so caught up in how big their innovation is that they stop thinking about the customer who is already successful and who sees the risk to implementing a model or technology. We don’t spend enough time addressing that problem up front. We’re too focused on the rewards.
___________________________________________________
Chetan has worked in the Software and Infrastructure space for more than 12 years and has started and run several companies in the space. As the CEO of the company he creates the vision for the company and its products and is accountable for the overall performance of the company.
|
