Since the inception of its first lab in 1945, IBM Research has grown to fill eight worldwide labs, employ about 3,500 researchers - including five Nobel prize winners - and help Big Blue become an IT innovator and market power across several technology disciplines. IBM's research executives credit the company's vast user community with helping the computer heavyweight determine product road maps.
Denise Dubie sat down with Robert Morris, director and vice president of IBM Research, to discuss how his team determines where to find tomorrow's technologies.
Q: Tell me a bit about what key technologies you're working on at IBM Research.
I run the Almaden (Calif.) research center. We work on nanotechnology, such as the very basics of our capabilities in physics and chemistry that allow us to be leaders in the microelectronics business. And we put those technologies together to form components. We also do research on how the components are assembled into systems.
In terms of systems, one of our major focus areas here now is on storage systems. We are building self-managing storage, which will continue to lower the cost of storage. Then on the next level up, we work on data management technologies, which of course includes databases, but it also means data mining and exploiting the Web. Above that we have a lot of work going on for Web technologies. Also we do work on the user experience, making sure that our systems interact well with people. Finally we perform research on how people interact . To really serve the services part of IBM's business, we have basic research going on regarding how people interact, how they communicate, how they collaborate and how we can improve the interactions between people as they solve IT problems. We like to say it's from atoms through to society.
Q: How does IBM decide what to research, what becomes a product, and what doesn't?
We are very careful to maintain a balance between free-ranging basic exploratory research and undirected research on the one hand, and on the other hand, research that is very intimately tied to customer and product requirements. We spend a lot of time making sure that we balance these. If you do just one or the other, you will fail.
If you only respond to customer needs, you will become an advanced technology shop, you will join up the dots and you will be ultimately unable to invent a disruptive technology. And if you only do basic undirected research, you will miss, because no research institution, no matter how good, will catch every trend. And even if you have those two, you're not guaranteed success. You have to fill in the in-between. There's a kind of a middle third, a middle area of pre-product. That's developing things that after a bit of research you realize and hope one day could become a useful technology or product.
Q: What has IBM recently passed down to the research group in terms of product direction?
The biggest activity is a really new mission within IBM of transforming information technology into an on-demand capability, which of course has a variety of aspects. Some people think of that as the making of IT available as a utility, and it certainly includes that, but it's much more than that. It's about transforming the whole IT operating environment. And over and above that, it's about transforming business.
Q: Where did the on-demand vision originate?
This is not something that is kind of a big bang. We're not going to say, 'This is when we'll launch our first and our last on-demand product.' If you go back a year and a half, one of the things we were talking a lot about was autonomic computing, and we're still working really hard on that. It is one of the first ways we began to discover this on-demand movement. CIOs, through hearing and learning about the technological requirements of autonomic computing, told us that that computing environment, the on-demand enterprise, is something they'd like to see.
It turns out that autonomic computing is the technology heart and soul of on-demand. It is the technology skeleton on which on-demand is built.
Autonomic has actually been going on for several years, and we didn't start on-demand several years ago. Our products had been component-autonomic even prior to that. We are now working toward being, to make up a definition, systems-autonomic or holistically autonomic now.
Q: Could you elaborate on how autonomic or on-demand research has found its way into IBM products?
To start where the human touches the technology, one of the most noticeable parts of that is using the client. In the PC arena, our ThinkPads and ThinkCenter desktops are well known, and within those we have created the autonomic client.
An example of autonomic is the way our systems are self-managing as they connect into the network. Our ThinkPads now ship with a technology called Access Connections, which automatically connects the client into the network that is there. You don't really have to reconfigure at any time. And new technology we're working on, which will come out fairly soon, will also automatically optimize your connection.
Q: What about security?
Naturally, that has to go hand-in-hand with security technology. We're working on a wide range of security technologies, including an embedded chip within our ThinkPad. That allows clients to store keys and do an embedded security-system capability so that all of the files can be automatically, on-the-fly encrypted when they are stored. So if you should lose your machine, the files are quite safe.
We can do a biometric authentication to our ThinkPads as well. And we make that available because we recognize that the way people connect with these very complex back-end IT systems is mostly through a client like a ThinkPad.
Q: How does IBM translate this type of security into wireless environments?
We are extending that technology now. Basically, the way you used to locate rogue access points was with a handheld sniffer, which you had to walk around the building with to find rogue access points. But now we've extended that technology to the point where every client, every machine, whether it be a desktop or a mobile, that is in the enterprise or on the campus, can participate in that. You don't have to have that sniffer anymore.
This is called Distributed Wireless Security Auditor. From any machine that's on our network it can triangulate, find rogue access points and then automatically go in and shut them down.
Q: How do user groups such as SHARE affect what IBM research develops?
It affects us in a couple of ways. Primarily it allows us to listen. Twenty years ago in IT you told people, 'Look at our new technology,' and then got their reaction to it. Today it's been reversed. Today you listen.
People in our company from the CEO down to researchers are out talking to customers from their CEOs to their network, database or storage administrators. It's what led to this on-demand movement. We just kept hearing from our customers that they needed to be able to work with a variable base instead of a fixed base because of the change in their industry. We heard that they had to be much more resilient toward disaster. Whether it was a market change or a disaster, we heard them say they needed to be more flexible.