Kevin Jackson is a luminary in the field of cloud computing. He is currently an engineering fellow at NJVC, one of the largest IT solutions providers supporting the U.S. Department of Defense. He is also editor of the Government Cloud Computing Journal and founder of "Cloud Musings" blog. Kevin was trained at the United States Naval Academy, Naval War College, and Naval Postgraduate School.
In this interview, we discuss:
- Actual government IT successes, especially using cloud computing
- Deconstructing the hype around cloud computing and talking about its real characteristics and value
- How clouds need a critical mass to function and provide elasticity, and how the size of the US places it in a strong position to develop these kinds of clouds
- Industry transition from system integration to service integration
- Use of clouds among defense departments and other organizations with other high security requirements
- Synergy between cloud computing and geospatial data
Robert Duffner: To get us started, could you please take a minute to introduce yourself and your experience with cloud computing?
Kevin Jackson: I'm the Director of Cloud Computing Services at NJVC. I've been in technology for a very long time. I started in the military, where I was an aircraft carrier pilot, flying the E-2 Hawkeye aircraft, specializing in command and control. I also worked with the National Reconnaissance Office, where I did a lot of early work in the delivery of intelligence products over global networks to mobile devices.
After I left the military, I focused on leveraging Internet technology in support of national objectives, eventually working with IBM in the WebSphere software group and developing SOA-based architectures to support the mobile devices that were coming into commercial use, developing infrastructures that delivered information to mobile devices securely.
That experience led to my interest in developing global SOA-based infrastructures for the government. Over the past few years, that came to be known as cloud computing, as it morphed into leveraging SOA on top of virtualized infrastructures.
I joined NJVC in February 2010, with a focus on the use of cloud computing to support the mission of central government, particularly the DOD and intelligence community.
Robert: It's always fun for the media to highlight government IT failures, and ZDNet even ran an article saying that Vivek Kundra, the federal CIO that you've had private conversations with, wants to get serious about Federal IT success. Can you talk about some of the big successes you have seen, especially in the context of federal cloud IT projects?
Kevin: One of the promises of cloud computing is that it allows you to leverage resources more efficiently. And the federal government is notorious in the inefficient use of technology, because each of the agencies and departments have built their own separate and independent approaches with respect to IT.
One area that the government has been underutilized in is the collection of data. The government has data on everything, from insects, to traffic, to where people live and where they move. That data has been collected for various reasons, but it is typically locked up in databases or storage, never to see the light of day again.
They store it because it's the public's property, but there is no impetus or resources to use the data for anything else. When Vivek Kundra first came in, he saw that as a waste of a valuable resource, and through usaspending.gov and data.gov, he was able to leverage commercial cloud computing offerings to set that data free. And by making this data, which is owned by the public, available to the public, new and amazing applications are being developed and released almost on a daily basis.
This approach was modeled after his experience with the D.C. government, where he did a very similar thing: Apps for D.C. So Apps for America, I think, is an awesome example of how innovation combined with cloud computing can provide extraordinary value. The army took that leap and did what's called Apps for the Army, which is a very similar thing.
They used the Defense Information Systems Agency's platform, called RACE, and offered that platform and its development environment to members of the army so they could use their innovation to develop applications that would support specific army missions. That pilot was also extremely successful, and I believe the Army is now planning on expanding it.
Their ability to leverage open source and to build a secure cloud infrastructure actually led to data.gov being hosted on the NASA Nebula Cloud, which was developed and built by Chris Kemp. That achievement shows that the federal government itself can be a successful cloud provider, which I think is huge in its own right.
Robert: That's definitely a great example, although alongside those successes, there's clearly an initial wave of hype, in this emerging market of cloud computing. There's broad-based skepticism about cloud really being all about hype, and about the government moving to the cloud being a boondoggle.
I'm actually thinking about an article titled "Government Incompetence Led to Compromise." It was on the AEON Security Blog, where the author asks why in the world did the government get suckered into the cloud? I think it was back in February. They talked about how hackers may have gained access through the content management system of third party vendor, GovTrends. I know there's also the Joomla CMS. Can you talk about some of the skepticism that you've encountered?
Kevin: First of all, I think there is a lot of misunderstanding around cloud computing. Many people I talk to think cloud computing is just the next technology du jour. They see it as the next marketing push; first there was mobile, followed by SOA, and now it's cloud computing. They believe that it's really nothing new, important, or different.
The first misconception is that cloud computing is a new technology. In my mind, cloud represents an evolution of technology that combines things like virtualization, SOA, mobile technology, and mobile infrastructures. It blends them together into a new model for consuming and delivering information technology.
Cloud computing is more a business model, an approach, and a technique than it is a technology. I think the first thing that's important to understand is that it's not new technology, but rather using and leveraging existing technology in novel ways.
Second, this is an inevitable transition that I sometimes liken to the industrial revolution. There, society went from an environment where products were handmade in cottages to assembly lines powered by the steam engine.
That was a revolution not just in society but in the economy and many, many different domains. What we're seeing with cloud computing is the exact same thing, except it's for information technology. It's the next step in the information revolution.
The hand-built, hand-designed works of art that we call our IT infrastructures are being transitioned to assembly line, modern, professional infrastructures composed of commodity components run in an automated fashion. So we're actually seeing the IT infrastructure being put on the assembly line.
I think it's going to have very similar effects to our society, the beginnings of which you can see today. My daughter is a freshman in high school, and she was born on the net. The way she needs, consumes, and reacts to information technology is completely different from the way I use and consume technology. I had to learn how to leverage the network. She was born into it.
So, that's what cloud computing is. It's not really hype, but rather a transition that we all have to be comfortable with. And I'm happy to see the federal government for once taking the lead in doing something that it really needs to do.
Robert: Across the board, we tend to be early adopters of IT here in the US, but government cloud isn't just a US phenomenon. You've pointed to the governments of Taiwan and Korea, in particular. How are you seeing similarities and differences in cloud computing requirements among world governments?
Kevin: When I was in Europe earlier this summer, I actually did a presentation on international governments and their use and adoption of cloud. One thing I found was that many governments are scared to death of the United States and its rapid adoption of cloud computing.
One of the key tenets of being a successful cloud provider is scale. You have to have enough scale so that you can leverage the differences between groups of users. That's how Amazon Web Services has been successful with providing infrastructure as a service. That's how Google can successfully get high efficiencies from its global IT infrastructure.
The criticality of scale, I believe, is one of the key reasons why cloud computing companies arose first in the United States, because you had a large market with a diverse user base that could leverage the capabilities of cloud. Arguably, the first cloud companies are Google, Amazon, and eBay. They are all US companies, and European governments see that as a threat to their own sovereignty.
Just take Belgium for instance. You may only have 10 million people in the country, but is 10 million people in a single time zone enough scale to leverage cloud computing? Will you ever see a Belgian cloud service provider? Take Africa, they may have leapfrogged other countries with cell phone technology because they didn't have to put in the wired infrastructure, but will that be possible with cloud?
And if United States companies become the largest cloud providers in the world, will that make the United States the key information superpower? These are the types of issues that governments have to deal with. It may be better to invest in having a domestic cloud infrastructure, even if you don't have the scale to be commercially successful.
If you're a central government and you need to provide services to your constituents, should you leverage a foreign government, a foreign power's cloud infrastructure because it's cheaper? Or should you invest in your own domestic cloud infrastructure so that your country can have that capability domestically?
Robert: In your session at Cloud Expo, you said that cloud customers must be able to easily store, access, and process data across multiple clouds. How are you seeing the interaction between private and public clouds evolving?
Kevin: I believe that in the delivery of IT systems, we're going to be transitioning from system integrators to service integrators, and these services are going to be delivered by multiple clouds. You have private clouds, public clouds, community clouds, and you may have industry clouds.
You will have services that are designed and optimized for specific verticals, and an information technology provider would need to understand not only what its customer needs are, but what services they will have to integrate in order to support all of their customers' needs. And it will be just too expensive to hire people to custom-design, build, and deliver these systems.
So as you transition from the system integrator role to the service integrator one, there's going to be a rise in requirements for cloud brokers. These brokers will need to understand the marketplace, including an in-depth understanding of customer needs and requirements.
They will need to be able to look at various types of clouds and figure out how best to leverage these services to meet their customers' missions, needs, and requirements. That's basically why I believe you will have multiple clouds and you will have to understand how to integrate these services.
That's another reason I believe the networking effect of cloud is going to drive an acceleration in the establishment of cloud standards. The value of the Internet grew as it expanded from a small nucleus to its global breadth today. That was relatively easy, since any node that wanted to join had to align with the standards that were already established.
As the basic Internet grew, it became more valuable. As it became more valuable, more people wanted to join it, which reinforced the existing standards. And that's why TCP/IP became the powerful global standard that it is today.
Cloud, fortunately or unfortunately, is starting with a global breadth from its very beginning. That means you're going to have multiple organizations trying to substantiate their way of doing cloud into the de facto standard.
The conventional wisdom is that it will be much more difficult to establish a standard, but I feel that the value of this new approach to IT will actually hasten the establishment of cloud standards. Cloud service integration and cloud brokerage will be an important catalyst to the economies of the world.
Robert: At our Windows Azure Customer Advisory Board meeting, we had Ray Ozzie as one of our speakers. He talked about how you have to have standards, and you have to have multiple offerings that appear similar to customers before you start to see broad adoption of technology.
Generally, I see that when your value props start to sound very similar to those of your competitors, it's generally a good sign that there's an understanding by your customers of what that offering is. It's clear we're not there yet with the cloud.
Kevin: The competition is going to be at the upper layers: at the serv ces layer, rather than at the infrastructural layer. I think at the infrastructure and the platform levels, we are going to see a rapid convergence of standards, and the competition will be associated with support offerings that are optimized for different verticals, industries, and tasks.
Robert: What do you think about the idea of a business process as a service?
Kevin: We will have value chains being created between clouds that are optimized for specific industry services. And there will be entire business-process management chains. Initially, they will arise internally, within enterprises as they look for ways to increase efficiency in their delivery to their customers. Eventually, they will morph into external, competitive offerings. In a lot of ways, salesforce.com is trying to deliver business-process management as a service.
Robert: To circle back a bit, you mentioned in a blog post that the army is taking cloud computing seriously, and you pointed to the Army Private Cloud RFP. What do you think is driving defense organizations in particular to embrace cloud technology?
Kevin: Government as a whole has an economic imperative to increase efficiencies and reduce costs. The DOD's mission is not to build IT infrastructure, but rather to support the goals of national policy, so the more they can reduce their IT investment requirements, the better. At the same time, though, they need to address the increasing need to leverage information technology.
The United States is arguably the world's only superpower, so we are seeing a transition to asymmetric warfare. The enemies of the United States aren't trying to out-muscle the United States. They're actually leveraging technology, speed, stealth, and the global reach of the Internet to force us to incur costs, both monetarily and in terms of loss of life.
Many of our enemies are against the concept of nation states, so this notion of warfare is going to continue, and the defense department needs to protect the United States as their core mission. Information is really the way that asymmetric warfare is waged against the United States. Cloud computing, quite frankly, addresses both the economic imperative of increasing efficiency and reducing costs while also enhancing our capability to fight this asymmetric information war.
Robert: You talked about the DOD investing in private clouds, but what are you seeing as the public cloud potential for defense and intelligence agencies?
Kevin: I see it as huge. Society lives in the open, and in order to understand what's happening in society, and the intercourse between nation states, the DOD and the intelligence community have to understand what's happening in the open society.
So, if it cuts off its ability to view, understand, and analyze what's going on in public, it basically does a self-imposed denial of service, and it will be impossible for these organizations to fulfill their missions. Understanding, leveraging, and participating in the discourse on the public networks is critical.
Robert: Clearly, you can have a very secure public cloud.
Kevin: Absolutely. I think that most of the reflex that there's not enough security in public cloud is misplaced. The vast majority of the information holdings of the DOD and the federal government is unclassified.
There are things that will never be in the cloud, but you simply don't put them there. Cloud is not a panacea for everything. It is great for some things and awful for others. The challenge is to understand how to use cloud for the things that it's good for, and also understand how not to use it for the things that you shouldn't.
And that's really all about in managing changing policies and procedures, learning more about the IT environment and the information revolution that we're all entering.
Robert: Let's talk about geospatial data. Back in June, Microsoft announced that SQL Azure, which is our cloud relational database server, now supports geography and geometry types plus spatial query support. It seems like geospatial data is an obvious intersection of government and cloud. Actually, you've talked about it and we're supporting it in Windows Azure. Can you talk a little bit about some of the use cases you see for geospatial data in the cloud?
Kevin: Since everything occurs in a specific place and at a specific time, all information has a geospatial coordinate associated with it. In order to understand the relationship of information to society, you also have to understand that geospatial coordinate.
I had a conversation a few months ago about healthcare reform. Apparently, it's a known fact that most healthcare fraud occurs in California and Florida, but a lot of the data associated with healthcare fraud does not have a geospatial linkage. So how can you design and apply national policies but also be able to understand when and where those policies should be enforced?
It costs money to enforce policies, so should you invest the same amount of money to prevent fraud in Florida that you do in Montana? You have a lot less fraud in Montana, so the money spent in preventing fraud there may have a far smaller return on investment than in Florida.
If you don't have a geospatial component to the data, then you have no way of deciding how to design, implement, or enforce public policy based on location.
Robert: I'm mindful of time here, so as you look five years into the future, what are some of the things you hope come to fruition with government use of the cloud?
Kevin: I would hope that more and more data will become available in the cloud, because with data, you get visibility and transparency. But you also get innovation, and innovation will drive society quicker. It enhances the economy, and as they say, a rising tide lifts all boats. If we can raise the economy of the world through the liberation of data and a more effective and efficient use of information technology, it will raise the status of individuals worldwide.
I see information technology enhancing the economies of the world and really enhancing society as a whole in the future.
Robert: That's a great place to wrap up, Kevin. Thank you for taking the time to talk today.
Kevin: It was my pleasure.