April 6, 2012
Kerrie Holley, CTO of IBM's Global Business Services division, sat down with InformationWeek last month to discuss IBM's focus, cloud computing, and the tech industry trends. What follows is an edited transcript of that interview.
With a background in mathematics and the law, Holley helped develop Service Oriented Architecture (SOA), a way to make enterprise software modular.
IBM Global Services CTO Kerrie HolleyIBM Global Services CTO Kerrie Holley
He holds several related patents, not to mention a patent on a way for mobile phone carriers to locate lost or stolen cell phones.
In 2006, he was appointed to be an IBM Fellow, the company's highest technical leadership position. Holley is one of 69 current employees in IBM's technical community of more than 200,000 people to hold that distinction.
InformationWeek: Let's start by reviewing where IBM Global Business Services is at the moment. How is the consumerization of IT, the mobile revolution, and the shift toward cloud computing affecting the company?
Holley: From a Global Services standpoint, there are three major focus areas that we have. Ginni Rometty, who's our new CEO, set the stage in 2012 in talking about us becoming an essential company. But if you look at a lot of what we're doing in Global Services, we do believe very strongly that analytics has a powerful opportunity to add a tremendous amount of value. In fact, we established a whole service line around analytics in terms of resources, training, capability. So analytics, I would say, is a big area, in predictive analytics and commerce and a variety of areas.
The second area is around cloud. We see cloud as more than just a paradigm shift. We actually see it as a business shift in terms of being able to preserve capital, being able to do business innovation, being able to actually do transformation, be able to actually do things that maybe business was heavily reliant on IT before but now has an opportunity to sort of shift that equation.
Then smarter commerce, and that involves a lot of what we're doing with retailers and others. It involves analytics. It involves cloud.
A fourth area is the whole campaign that we've done around smarter cities.
InformationWeek: Does IBM's focus on predictive analytics involve a compliance practice? There's a lot of focus on privacy regulation at the moment.
Holley: We focus not only on the infrastructure but also on the business side in terms of what you can actually do to drive better business outcomes. In terms of security and privacy, the answer is yes, we do focus on that. We focus less on the compliance issues that maybe an auditing firm might spend more time on, in terms of whether or not you're violating a law in Croatia. But yes, a very strong interest of ours is to ensure that we're not doing anything that would undermine the credibility of our clients or ourselves in terms of privacy and security. We actually have a privacy and security practice in this space.
InformationWeek: What really excites you the most in terms of what's going on in technology now? What can you not get enough of?
Holley: It's a great question. I think the area that excites me the most is that I've spent most of my career working, to be honest, extensively with IT, with a heavy business focus. So, often the CEOs or other C-suite executives have imperatives they could look upon IT to facilitate and [to] adopt. One of the biggest challenges that businesses have been facing is to get IT out of the loop. When I say "out of the loop," I don't mean to eliminate IT, but to not have IT as a bottleneck, whether [that means struggling with legacy systems or waiting on service queues]. [To put it another way,] I'm actually very focused on right now is this whole concept of business agility. How do we actually create outcomes faster that matter? How do we actually make agility something less than a platitude, something that sits in a strategic document as "We want more," and something that can actually be engineered?
So what excites me is that we have a wealth of technology today that's come together to allow us to create business-centric approaches, where the actual business is actually driving the approach and the method and the adoption, where technology is being leveraged to actually drive these better outcomes.
InformationWeek: Can you provide an example of that?
Holley: I can. [This involves a project we did in the Eastern part of the world for an insurance carrier, working with a telecom provider.] I'll contrast two scenarios. One scenario is where I'm looking to be maybe number one. "I want to be number one in my market, but today I'm number three." So a traditional approach might be to attack the problem by looking at, "OK, maybe I'll grow more customers. Maybe I'll offer new, innovative products." Not that those are bad approaches, they're great approaches. But at the end of the day, the outcome of being number one may not have been achieved. And you may have met your metrics. You may have actually achieved more growth in clients. You may have achieved more differentiated products. But perhaps your competition did something that was equally as productive, and you were already behind.
So the former approach focuses on this notion of improving processes and efficiency, whereas an agility approach would not solely focus on that but would actually look at, "OK, what do we actually do?" So first of all, it approaches on the outcome that you want. The outcome that you want is not to simply be better. The outcome that you want is to be number one. So in order to achieve that outcome, you have to actually identify imperatives that can be achieved in your window of time, which, in this case, was literally a 12-month window of time, that would actually propel you into that number one position.
So, first and foremost, [the business agility approach] focuses on the key performance indicators that must be achieved. That's sort of obvious. But then secondly, it requires this fusion between how the two teams work together in terms of what innovation is necessary versus what's on the truck that we can do right now.
I can give you more detail, but that's the base difference in approaches. In the second scenario, we don't just look at the processes and try to make them more efficient and more effective. Instead, we look at the agility indicators in terms of outcomes that would actually make a difference.
100 Years Of IBM: 25 Historic Milestones
Slideshow: 100 Years Of IBM: 25 Historic Milestones (click image for larger view and for slideshow)
InformationWeek: How is the shift toward mobile computing and different form factors of computing, like tablets, affecting what IBM does? Has it shaped the way the company goes to market or the direction the company is headed?
Holley: I would say a couple of things. One is that mobile computing is certainly huge, and a huge opportunity. We have practices, both in software and research and services, focused on that. From a Global Business Services standpoint, or maybe from a market standpoint, there are a lot of boutique firms that will offer capabilities [that may be more specific with regard to mobile devices]. What we do, from a services standpoint and from a software and research standpoint, is we have a user-centered design team. We have a group of people that specialize in the user experience ... to actually create a rich experience on a mobile device.
Then there's another part of this equation, which is to actually maybe create a rich mobile experience. We've done a project in India, and this was work done out of our research, where we actually created voice-enabled Web. That's an example of offering a capability that hadn't been offered before.
In fact, I'm beginning a project at a major transportation company, where the thing that we're trying to do is ... [to allow customers to configure vehicles being purchased] using a mobile device, where you can actually customize your vehicle. At first glance, it seems rather simple. But in certain types of vehicles, the configuration options are mind-blowing. So we actually have a research asset that we created which actually helps this decision [by using algorithms to winnow the options] so we can actually present the information on a mobile device. A lot of what we do, arguably, is on the back-end side, but it's specifically geared toward the interface, which we know will be mobile. We know it will be a smartphone or a tablet of some kind, versus someone sitting at a desktop.
InformationWeek: Is there any particular challenge or problem facing the IT industry that you wish could be changed?
Holley: Absolutely. There's probably a few problems of that kind. As you know, we're a B2B company, so largely the problems that we describe are problems that major businesses, or even small businesses, but businesses that use technology face.
I'm working with this company in the South--and it's not the only company--one of the huge problems we see is that a lot of applications that are built in the 21st century turn brittle within a three-year window. What I mean by brittle is the ability for the business to actually leverage that application for future opportunities, whether it's a new process change or a new market opportunity. Companies find that they're being constrained by systems that they built, [some 20 years ago and some as little as three years ago].
That's one of the reasons I spent a lot of time with Service Oriented Architecture was to [avoid that type of problem when we] actually build business systems. So I would say that challenge is still upon us, but it's certainly one I'd love to see change, because I think it would free up capital, it would free up resources, and it would allow businesses to spend a lot more time on new things, versus keeping the lights on and integrating things.
InformationWeek: By Service Oriented Architecture, you mean a more modular approach to the design of systems so that they can be plugged into new systems without ripping everything out?
Holley: Exactly. At the heart of service oriented architecture, it builds in flexibility into what you build, so that you can do exactly what you just described--begin to treat these systems as LEGO blocks so that I can take a piece and it's not constrained by the fact that was attached to this back-end or attached to this front-end, and I can move it to a new usage without a cost.
InformationWeek: Technology companies have long relied on technical incompatibilities to protect their markets and gain a competitive advantage, even as they have to make some concessions toward interoperability. Companies have to find their own level of comfort with being open and being closed. Do you see this balance shifting?
Holley: I do see it shifting. If you look at the physical world in terms of hardware, we can argue whether this is good or bad, but you and I, when we no longer have a device that meets our needs, we just go buy a new one. Or if it breaks, we go, and when it's repaired, it often means getting a new one. But when you think about that, that's because the cost of doing so is rather low. But when you think about software, the amount of time and money it took to build some of these systems unfortunately costs a lot more than a smartphone and costs a lot more than a refrigerator or these other devices.
But at the same time, we'd like to move to that same model. We'd like to be able to discard and replace. I think what you said is true. I think that some of the cost [of incompatibility] is a necessary evil. But I do think a substantial part of the cost is how we actually built it in the first place, and I think that part of the problem we can fix. We find that's being fixed as we speak.
In fact, this company, which is a transportation carrier that I'm working with in the South, this is exactly what they're going to do. They're not the first ones to have done this, to have taken old systems and revamped them. The reason they do that ... is not repeat their problems or sins of the past. They don't want to redo it and then find that they didn't actually build in this flexibility as an attribute of the actual new system.
InformationWeek: In terms of software design, it could be something as simple as just making sure that there is an XML export capability to ensure forward compatibility or something like that?
Holley: Exactly, that's a great example. We use this term of "agility indicators." But there's a ton of these indicators, or rather, attributes, these levers that you can pull. The more that you pull, the more your resulting system is more agile. But the XML is a classic example. Web service is another. But there's a bunch of these rules, the externalization of business rules. There are a lot of these levers, when applied holistically, have a tremendous benefit to the outcome.
InformationWeek: In cloud computing, there are trade-offs. You get a lot of freedom in terms of not having to worry about systems and not having to worry about the scalability of systems. At the same time, there's concern about lock in, security, and future-proofing and so on. How would you describe IBM's view of where the cloud needs to go in terms of competing with the advantages of on-premises systems?
Slideshow: IBM Empowers Smarter Cities
Slideshow: IBM Empowers Smarter Cities (click for larger image and for full slideshow)
Holley: Another excellent question. First and foremost, there are these types of clouds you can pick and choose from--public, private, or hybrid, or community.
Let's say I have the challenge, which you describe, where I'm concerned about lock-in, I'm concerned about data separation and security. So the issue, I would argue, is not a cloud issue. It's an age old issue of, "OK, how do I avoid the lock in? What are my actual concerns?" Maybe my concerns are that I don't want a strategic asset to be cloud-enabled. Or if it is cloud enabled, it's going to be a private cloud, so maybe I'll avoid a public cloud. It's an example of a decision point.
Or if it's an issue of data security, maybe I'm going to go with a vendor that's got a proven track record. Maybe I'm going to go with a vendor that can guarantee and assure me that they actually can do this [data] isolation, this segregation. And maybe more importantly, I'm going to have an adoption strategy that has a transition plan that allows me to actually shift from the cloud vendor back to my own private cloud if I choose so.
In fact, a colleague of mine, we've actually got a book that's going to be published soon, "Is Your Company Ready for the Cloud?" that actually goes through some of these decision points.
Our perspective is that it's not so much an issue of cloud computing as it's an issue of your adoption strategy and your decision framework.
InformationWeek: Let's talk a bit about your own role here as CTO. What's your typical day like? What do you really have to oversee on a careful basis?
Holley: My Sunday evening is probably a good example, where I'm on the phone with China, I'm on the phone with some folks in the U.S., because of a strategic project that we're about to launch. We spend a few hours just going through the project, going through what are the things we need to do on Monday to actually get the project off to a great start, what are the risks, how do we codify and analyze those. That's not atypical.
Then, if we look at today as an example of Monday ... I'm again on the phone with some folks in Toronto, some folks locally, and some folks in China, where we're looking at our asset strategy. Like any organization, we have funding challenges, we have maintenance challenges. The question is, how do we keep these assets vibrant so that the teams that need them can [use them] ... How do we ensure that this asset is maintained consistently with our software products? So that's another issue.
Inside Watson, IBM's Jeopardy Computer
(click image for larger view)
Slideshow: Inside Watson, IBM's Jeopardy Computer
Holley (continued): Then, today is not atypical, where I write. I'm finishing up some chapters on this book. I was also on the phone about a week I'll be spending up in Toronto for a project that we're going to launch in South America. We'll spend a week with our Smarter Planet software colleagues and some research colleagues, and we'll spend a week looking at the assets that we want to apply and when we want to apply them, how do we want to apply them. We'll look at the architecture and describe how we can actually minimize some of the risk. This is a huge project, and it'll be fairly significant when it's publicized.
InformationWeek: Are there any insights from your cloud book that you'd care to share?
Holley: Yeah. I think one big insight. Pamela Isom is the primary author, and I'm a co-author. But Pam and I, what I think is most useful about the book is this notion of a cloud adoption strategy, which goes to the heart of [what we've been talking about]. We spend a good amount of time in the book describing such a strategy, how it's linked to your enterprise architecture, and why those two are important. We offer up some decision frameworks. What the book doesn't try to do, it doesn't try to be a primer on cloud, it doesn't try to educate people on cloud, although some of that's in there. But instead, what it tries to do is help companies that are beyond those issues with making these key decisions and what the actual impact of those can be. We also try to focus on looking at cloud computing as more than just a cost-efficiency issue. It can actually be used to drive innovation and make a business even richer in some of its outcomes.
InformationWeek: Are there any examples of that that come to mind?
Holley: When you look at analytics, as an example, it's very expensive for some companies to invest in the information supply chain necessary to make analytics really pay dividends for them. Suddenly, with the cloud, that capital preservation is possible, yet they can, at the same time, take advantage of this analytics to drive innovation that maybe they otherwise didn't have. So I think that's one example of how the use of cloud is an enabler to do something they couldn't do before.
InformationWeek: What about your background in technology? You started off in mathematics?
Holley: I did. InformationWeek: Do you come to technology primarily from a math background, or did you also get into the engineering side of things?
Holley: It's a bit of both. When I entered the marketplace, the computer field was not as ubiquitous as it is today. So in high school, actually, was where I learned how to program. Even though, in high school, you have a general purpose program, I actually started programming as a sophomore, junior, and senior. Then when I went to the university, I predominantly did mathematics, where computer programming was sort of ancillary as a complement to drive and solve certain math problems. Then when I left the university is when I focused entirely on software engineering.
InformationWeek: What were you programming in?
Holley: A variety of languages. At the high school level, it was BASIC and Fortran. At the university level, it was Pascal, it was some C ... those were the dominant languages at the university.
InformationWeek: I take it you don't have much contact with coding today, or do you keep up with it?
Holley: I don't have the same contact. But I do keep up with it. ... The dominant computing languages that we use today, I'm pretty fluent in--Java, COBOL, and C++. But at the same time, as a [software] architect, I don't have the [time], nor is it cost efficient, and probably I don't have the skill to be as fast as some of our younger talent. So I'm aware of what works and doesn't work. If we had to build something and we needed something to throttle, I would know about what is necessary in terms of exception handling. So no, I don't actually program anymore. ... But I keep up with the patterns, design patterns and architectural patterns, and assets that we have to build using these patterns and frameworks, open-source frameworks, non-open-source frameworks, to drive adoption and accelerate development.
InformationWeek: Do you follow it in the sense that you have a sense of where programming is headed?
Holley: Yes. [One of the areas that is interesting to me is this notion of data scientists, people with] this data intensive computer science background, where they actually know how to program, in languages like Python or R ... to extract insights from data. So my focus is more on the former in the sense that I'm keenly aware of the kinds of technologies programming languages being a type of technology that we need to apply to solve certain types of problems, and how we recruit, how we train, and where we should be focusing our energies in terms of talent. This is an area that I think will grow tremendously over the next several years.
InformationWeek: Is our education system producing the kind of technical talent IBM needs?
Holley: The talent is definitely there in the United States. Whether we have enough of the talent, I don't know the numbers, but clearly there are a lot of people who believe that we have a shortage. I won't deny that. That may be so. But there's certainly a lot of talent in the United States. ... I would grant you that we could have more. But one of the things that our universities are doing well ... is this notion of a scientific method in oral and written communication skills.
I think that's where [U.S. students] have a tremendous advantage. When they can combine that with the mathematics and the engineering skills, then I think we have tremendous competitive advantage over some other countries. But at the same time, we recruit all over the world, and we get a tremendous amount of talent from all parts of the world, where there's people being trained in doctorates in mathematics and other physical sciences that are a great advantage to us.
9 Startups To Watch In 2012
9 Startups To Watch In 2012 (click image for larger view and for slideshow)
InformationWeek: I'd be curious to hear your assessment of the patent situation in technology, because it's a very contentious one, and some people have very differing opinions about both the patent time period being appropriated across all industries, and also the extent to which patents should be applied to things like software and whether the terms should be more flexible.
Holley: It's a great question. We were one of the big proponents of not doing so many process-based patents. So I think that was a good choice and a good decision because patents, I think, should be about things that we can physically touch and see, software being something we actually can sort of touch and see. At the same time, I think that innovation should be rewarded. When people create things that are new to the marketplace, I do think that they should be rewarded.
That's just my philosophical perspective. But I do think, at the same time, I know that we have a lot of joint agreements with companies who are both our competitors and our partners, where we jointly license our patents to each other, and I think that's the right decision to make our market more healthy.
I think that most companies would agree that companies that are strictly in the business of being predatory around their patents is not good for the industry, that what we really want to do is not to stop growth and innovation but to encourage it, and so I think the cross licensing is a great way to accelerate that.
In terms of my own patents, you know a lot of my patents for the general market would be perhaps obscure might be the word. A lot of them are in the area of SOA. One that I find interesting I only came up with it because I was losing my cell phone constantly and frustrated with the fact that my carrier couldn't help me locate it. So I came up with a patent on how to identify, locate, and secure a lost, high-value mobile device.
One of the ones I'm also pretty excited about is something we just filed recently around business agility.
InformationWeek: What do you think about the patent term being the same across industries? Some people argue that a 17-year term may be appropriate for a pharmaceutical patent but not for software, given how quickly the industry changes.
Holley: I'd say it's a great point. I actually have never thought about it before. I think there are some differences I would argue so I could see the argument to why the period should be shorter, but I would also be cautious. I think one of the reasons you have this time window is because sometimes the ability to apply the patent--and maybe that should be the test versus a window of time--is because sometimes the patents are ahead of the other innovations that will use them which is not true in the pharmaceutical industry, right?
Those patents will be used and applied immediately, whereas in the software field we might find some patents ahead of their time that have actually created something that could be used but it may be four or five years before other things are in place. So I think it's a bit more complex ... I can see both sides.
InformationWeek: Who do you see as competitors at the moment?
Holley: Well, we compete with the big companies like Accenture. We compete with a lot of boutique firms as well, but Capgemini, Accenture, so the whole legacy of the whole big six are still our competitors. ... We have to offer value propositions that make sense, that offers the IT department and the business a better alternative. We're sort of competing with ourselves. We're constantly trying to offer opportunities for them to do better with us than they could do on their own. InformationWeek: Are there any internal IBM projects that you expect will have a major impact on the tech industry in the next few years?
Holley: Yes. I spent some time with the Watson team. We built a team of about 100 people. My former boss runs this team, and I think the potential of where we can leverage natural language processing, this subset of artificial intelligence is enormous, [particularly in the] medical field and the banking field. That excites me a great deal in terms of what I'm seeing that we're able to do. We also have a lesser-known technology called "Deep Thunder."
Deep Thunder is all about doing more with being able to engineer the output of the effects of weather. [We do a lot of weather forecasting but] we don't say, "OK, we know within 24 hours this storm is going to hit Rio de Janeiro, and it's going to hit this particular district." This has a tremendous opportunity if we look at what happened in New Orleans: The potential to actually deploy resources to the actual areas that will be most impacted I think is a pretty interesting technology.
InformationWeek: What is IBM doing to make governments aware of such technologies?
Holley: The whole smarter cities work that we're doing is really not just a pro bono activity. It's also a way of helping cities across the world of understand what's possible with technology. When you look at what we're doing with the smarter cities program, each of the cities has a very unique problem. Some of them are very unsophisticated in their use of technology, and some are very sophisticated, but in each case there's something that we can bring to the table to make that city smarter. I mean the whole reason we did Watson in 2011 was to make the world understand better what technologies [are] actually possible.
I didn't mention that but that's another part of what I do is I do speak at conferences. I did this whole presentation on the future of analytics, and a few executives out of Sacramento State who happened to be present ... got really excited on the possibilities because they saw the connection between their problem and the vision that I said is actually possible. So now I'm driving up to Sacramento tomorrow to have a more detailed discussion with them.
With any client who is trying to do something out of the box--trying to build a house or building that stands out--the architect is the one who helps the client see that vision of what's possible.
In our industry, I see that as part of my role as a software architect and as a CTO, to help businesses see the art of what's possible with technology in a very like manner. InformationWeek: What can technology do for companies today that you'd say is really surprising?
Holley: Well, I would say we can do some enormous things. We can actually look and triangulate feeds. They're public, from social media. We can actually look at data in the actual stores in terms of behavior, and we can look at competing data across a marketplace and be able to actually advise a store on what they ought to be able to do differently. Something as simple as telling a retailer, "Hey, did you know that by doing returns, even though someone bought the item online, if you allow them to return it at the store, it results in X amount of dollars more sales for you per transaction?" That's a small insight, but it's a very profitable insight.
Or in the City of San Francisco, "Did you know that 70% of your traffic gridlock is because people are searching for parking, and well, what if we were able to take that window of time of search down by five seconds, do you understand the impact it would have on your city traffic?" I mean I think that's another powerful insight. So there are a zillion of these insights that we can do with analytics that requires not only the technology but the people to actually go out and discover these patterns that they would have never thought of before.
I think a new era is upon us. This new era clearly leverages the Internet but I think this new era ... I mean we've called it smarter computing ... everyone has a different name for it. But that's not so much what matters. What matters is, are we correct? Will we see a spend increase in terms of the percentage of gross domestic product? Will we see a spend increase over IT?
I think we will. I think what that indicates is that businesses are going to believe that they can actually gain more performance by spending more on IT, but it's not going to be on the traditional things. It's not going to be on building better warehousing systems or better management systems.
It's going to be on this fusion that we see between as we call it, "the world getting smarter." But it's going to come from leveraging business and IT applications. It's going to be out of leveraging stuff out of the social media. It's going to be out of leveraging things we actually see, physical sensors and robotics and GPS, and geospatial data.
The pay-as-you go nature of the cloud makes ROI calculation seem easy. It’s not. Also in the new, all-digital Cloud Calculations InformationWeek supplement: Why infrastructure-as-a-service is a bad deal. (Free registration required.)
About the Author(s)
You May Also Like
The Era of generative AI-enabled Security
How to Develop an AI Governance Program
Solution Brief: Fortinet FortiFlex Delivers Usage-Based Security Licensing That Moves at the Speed of Digital Acceleration
The New Frontier of Cyber Security: Securing the Network Edge
Top Six Recommendations to Improve User Productivity with a Hybrid Architecture