IBM Research is a 60-year-old institution that's had its hand in many technologies people now take for granted. The list extends to everyday devices like hard drives and programming languages. But today, about half of Big Blue's revenues derive from consulting services--hardly a hard science.
Horn's job is to make the squishy field of services resemble something like an engineering discipline. Without applying technologies and research to services, IBM Global Services risks being undercut on price by lower-cost rivals.
He also has to strike the right balance between fundamental research in fields like nanotechnology and the needs of IBM's product groups, which count on inventions to get an edge over competitors.
Horn spoke to CNET News.com about research in services and IBM's investment in emerging technologies, from natural language processing to special-purpose hardware appliances to open source software.
Q: Microsoft has a comparatively sized research organization. How would you compare IBM and Microsoft Research?
Horn: We have a very different corporate culture and corporate focus than Microsoft. IBM would like to think of itself as sort of a major player in the open communities. We really think of ourselves as helping others use our , helping open communities within which we operate and flourish.
Microsoft is sort of where we were at 15 or 20 years ago. They are manically focused on how they will build their next generation of their application suites and how they will maintain the Wintel duopoly and how they will maintain Windows position. That's perfectly understandable given where their company is, but we have a completely different focus
Many disruptive technologies emerge that can threaten established businesses within IBM. How do you get these into the company?
Horn: I spend eight hours with (CEO) Sam Palmisano and the senior VPs of the company every December to discuss the disruptive technologies that are coming down the pipe. I'd like to think that that presentation I make every year and the follow-up action, at least in the recent history, helped us avoid missing major disruptions. And there are major disruptions that are coming down the pipe. It's interesting to watch some of our competitors when they hit those disruptions, it's much more traumatic for them than for us because they haven't seen them coming.
Let's take an example. What company would you say would fight open source the most? We have over a billion dollars a year of bottom line profit from licensing intellectual property and yet we're embracing open source because we know it's going to be completely disruptive. We are the biggest scale-up (server) manufacturer but we built two-processor (servers) and we builtHow do you split up your research dollars? (a parallel supercomputer). So we're ready to embrace those disruptive technologies. It's not easy but it's better we do it than someone else.
Horn: We spend in the ballpark of one-third on services, one-third on software and one-third on systems and technology. Our portfolio in research investment is not exactly one-third and one-third and one-third, but it's been moving from hardware, which at one time was 85 percent of the division, to now more software and services.
How does research play into services delivery?
Horn: In services, to a very large degree, we will take over someone's data center. Or we will provide applications that solve problems for customers and we'll provide them remotely. And people pay for that as a service.
So you can go after this business in two ways. One is, you can basically just do labor arbitrage, which means you do whatever you can in India or in some low-cost labor environment. Or instead of lowering the cost of the people, take the people out--use automation. Research is really good in helping actually with both of those but really good at helping with automation.
Sounds like you're investing in technologies to make your consultants more efficient and effective in engagements.
Horn: It's the same thing that we've done for years in manufacturing. You use automation in manufacturing to fundamentally lower the cost of manufacturing, when you take out as much physical labor as possible. This is the same--except that technology hasn't been applied to IT services delivery. So there is more opportunity for automation, and the surface is barely getting scratched.
Palmisano has touted this idea of process improvement, where your business consultants can analyze a business process by breaking it down into its component parts. Then you can tell them how best to streamline their processes. Can you explain?
Horn: That piece of the business tends to be more in consulting practices and in what we call Business Performance Transformational Services. This is really one of the single biggest opportunities for IBM.
Think about it this way: Companies have some inherent inefficiencies in the way they work, and they do certain processes themselves which they could potentially get from the outside. Or they don't integrate external processes and services particularly well with their internal stuff, and all that leads to inefficiency.
Just think of the gross domestic product of the world with all these companies operating in some inefficient manner. If you could squeeze a little inefficiency out by detailed analysis and modeling of how a company operates--there are huge opportunities there. By the way, you can be very innovative and create a new iPod and you're talking about tens or hundreds of millions of dollars. And here we're talking about hundreds of billions of dollars, so the opportunities are really big.
What are some of the things you're excited about in hardware?
Horn: BladeCenter. That's a huge, huge opportunity. One of the big research projects (here) is exactly in that space: to make it simple for people to just plug in accelerator boards, and your application will run faster. No software, no nothing--hands off by the customer except one plug, and all of a sudden the whole system runs faster. That's harder than you might think.
Horn: Absolutely. Actually, a lot of DataPower's strength is its clever use of software. So it's not really that it's unique hardware as much as it is unique software. It is applied in a box or in an appliance that you can just plug in, in the front of your servers, to speed up your XML processing. You're going to see more and more of those sorts of things.
There are a lot of underlying reasons why that's going to happen. Traditional silicon technology is running intoand other problems. There are a set of things that are going to make it more attractive than ever before for having special purpose (appliances). There have always been accelerators. The problem is that by the time you program the damn things, the general-purpose processors got better. So you had all applications running on general-purpose processors and except for niche markets, accelerators never made it. What's going to happen now is that the general purpose ones aren't going to accelerate so fast, and the technologies and tools and software for integrating the accelerators have gotten better. So you're going to see a lot more special purpose stuff that you can just plug in simply.
One of the high-profile projects that came out of IBM Research in the last couple of years was
Horn: It continues to be a big thing for IBM and for IBM Research, but it's not just WebFountain. The basic issues are, really, natural language understanding in general. What WebFountain was able to do, which made it powerful, was it would go in and would scan text documents on the Web and it would understand enough about what people were saying that you could query it about what people were saying. You could imagine that there's a lot of countries, including our own, that would care a lot about scanning documents and even open documents and crawling through them to see what people were saying. A lot of the early work on WebFountain was done in three languages--English, Arabic and Chinese--and you can guess who might sponsor that work.
WebFountain is an example of a natural language technology that allows you to essentially analyze from an intelligence point of view what people are saying, but the important point is that this is just a small piece of many, many problems that companies have and where you want to take advantage of natural language understanding, such as translating spoken English to Russian and back again.
We talked about call centers. Natural language understanding can be incredibly powerful, even if you've got a call center operator, just by monitoring the calls and trying to understand what the issues are. There's enormous amounts of natural language and analytic issues in how companies interact with their customers. WebFountain was a specific application of natural language and search technology, but it's just one.
It sounds like you're very involved with customers, and IBM Research Center makes a lot of money from licensing. Do you deliberately try to research commercially oriented products attached to specific product groups? How do you manage that process?
Horn: It's tricky, let me tell you. We try to keep a balance. We don't want to be like the old Bell Labs, which was exploratory research but no connections to the marketplace. We want to make sure that we build channels for the flow of our innovation and our ideas into the marketplace. We spend a lot of time thinking about getting the right balance--that is, you can't just be doing short-term market work and short-term development or you become a development lab. You can't just do corporate funded basic research or you don't have the channels for the flow of your intellectual property into the marketplace. So you need a balance, and everything we do in the division is focused on that balance. In the end, to me that's the big trick of running a research division: getting that balance right.