Intel Labs Advanced Research and Development, Q and A with Kevin C. Kahn
United States Home | Select a Location
Site Map | Contact Us | About Intel
Advanced Search
Home ComputingBusinessDeveloperReseller / Provider
Intel Research and Development
Technology Focus
Silicon Technology and Manufacturing
Microarchitecture and Circuits
Computing Platforms
Software Technology
Communications and Networking
Technology Highlights
Expanding Moore's Law
History of Innovations
About Intel R and D
Research Links
Intel Research
Microprocessor Research
Technology and Manufacturing Group
Strategic CAD Labs
Global Manufacturing and Supply Network
News and Press
Intel Developer Forum
Intel Developer Update Magazine
Intel Technology Journal








Intel Fellows, the company's highest technical position, are selected for their outstanding contributions to the company. This career path allows an engineer to focus on and become an expert in one or more specific fields or areas. Intel Fellows have shown that they can exercise influence internally and externally in the industry through their competence, significant technical papers and patents, as well as external awards and honors. They serve as effective role models and mentors for other technical contributors in the company.

Q1: What is Intel's focus in communications and networking?
A1: That's a hard question because we focus on all levels of the technologies that enable data communications. Let me define what I mean by all levels. Our focus begins at the physical level, whether it's high speed Internet, phone line data, phone networking for the home, optical switching, or radio waves. It ranges above that into the communications protocols, TCP/IP, and all the basic Internet protocols. Then there's communications middleware, things related to how communications devices synchronize with one another, how they roam, how they bill for roaming, and how they keep people in connection with one another. And there are the infrastructural platforms such as APIs, how you expose communication capabilities to pieces of software. So it is this whole package of things. Interestingly enough, communications is a unique area at any company, in that it spills over into the regulatory space.
Q2: What is it about the regulatory space that is unique?
A2: Most of the technologies we deal with do not primarily impinge upon any regulatory issues. Communications is quite the opposite. There is a long history of regulatory involvement in communications at national and international levels that you have to be cognizant of if you are going to operate in the communications space. So we work in some of the public policy spaces to make sure those technologies are viable from a regulatory perspective by the time they show up in the real world.
Q3: If I asked you to give me the classic elevator pitch explanation of Intel's focus in communications and networking what would you say?
A3: I think at the end of the day, the thing that ties it all together is we want any smart electronic device to be able to communicate with any other smart electronic device wherever it is using the best available communications channel it can find. And to make that as invisible and seamless as possible so that people do not have to think about it.
Q4: Can you give some examples of key research and specific projects that may be going on now that are an example of this direction?
A4: Sure, lots. Let's start with the physical technology. The two areas in the physical layer that are getting the most emphasis are optical and radio. We have actually had a long history in other kinds of physical layers. For example, we did a lot of work in both cable modems and DSL technologies a few years back. Those have both now rolled out and are widely available products for various people. Today I think the hot spots are radio and optics. So, in the radio space, for example, if you want to talk about some pure research exercises, things that are pretty far way from product, there is an interesting technology called Ultra Wide Band. It looks potentially promising as a very high data rate short-range communications mechanism. It is a technology that has only become possible as signal processing capabilities have become much more powerful. It is a technology in the radio space not because theoretically it is new, but because the processing power to even comprehend trying to do it that way is only now with us. We have got a research project going on in Ultra Wide Band right now.
Q5: Anything else in the radio space?
A5: We are starting to look at building affordable radios and integrating them into other electronic devices. We are looking very hard at how to build radios using seamless technologies. There are a few people out there in the market working with these types of technologies, but we would like to be able to figure out how to do it using Intel's high volume manufacturing processes. Ultimately, it would be very nice if radio could simply be yet another functional module that you could put on a chip when necessary. We are quite far from doing that for lots of reasons how do you integrate what has been traditionally analog processing into a digital-type of process? How do you run a radio which fundamentally deals with very, very weak signals side-by-side with a processor that may be running at multi-gigahertz and generating a lot digital noise? How do you isolate these things? Is that possible? Is it cost effective to try to do that? There is also some research going in what is technologically a fairly distant space in many ways Micro Electronic Mechanical Systems (MEMS). There are certain applications of MEMS that look like they might make radios more cost effective. So we are looking at if we can apply some of the MEMS technology to our kind of environment and our kinds of radios.
Q6: What about wireless?
A6: There are a lot of standards activities going on in wireless. We are particularly focused right now on trying to sort out what will happen in the five-gigahertz wireless LAN band. There are at least two standards in the national arena 802.11a and ETSI HiperLAN2. This is an example of where I said regulatory issues tie in. For example, 80211a, an IEEE standard, is not legal in Europe. We are working on changes to the standard as well as lobbying with the European technical community in the regulatory space to work on making sure that this technology is legal in Europe when and if it is available. We also have people looking at a lot of issues in cellular 2.5G, 3G, and beyond.
Q7: And on the optical front?
A7: There we actually have research projects going in optical switching that are looking at cost effective ways of building optical switches using Intel's type of process technology. What are the limits there? Again, this is an area related to MEMS. You can look at MEMS as a way of doing optical switching.
Q8: What about the PC? How does that fit into your communications and networking research?
A8: We're looking at the whole environment, the extended PC and how it can communicate with devices around it. There's Universal Plug and Play (UPnP), a uniform specification for how devices find each other on a network, particularly in the home, but usable in a lot of other environments as well. Looking at how devices communicate out of the home how does a PC get through a residential gateway, how does it find a gateway, how does it configure the gateway, how does it do all of that automatically? Obviously when you talk about using this technology at home, people are not experts by any means. Probably their 14-year-old kid is the best administrator they have got. We need to make that technology just work. We are working on technologies that will allow this to happen transparently to the user.
Q9: What about wireless and radio for the home?
A9: That's something we're looking into too. Efforts like how to carry video over radio in the home environment? 1394 is an IEEE standard for multimedia communication that has been embraced by a lot of consumer electronics companies as a way to communicate video, say between a video camera and another device. But what if I want to carry that wirelessly? How do I do that? How do I do it over an 802.11-type communications channel? It's tricky to carry things over those kinds of channels. It requires certain quality of service enhancements to 802. We are working on those. Specifically, we are engaged with the 802/11e working group, which is working on quality of service for 802.11 radios.
Q10: A major thrust for Intel in recent years has been networking. What's happening in the labs in basic network technologies?
A10: We are one of the companies encouraging the notion of network processors. These would be dedicated processors that operate at the speed of communications lines speeds on packets as they come through the switch or router. They would operate very quickly on a particular type of computing application, and cooperate with higher level processors in the system that can handle more complex tasks that are only required on a very small number of packets. To make a network processor available, you need standards, so there is a lot of effort in what we call network building blocks. We helped create the Network Processing Forum, which is an industry group that is addressing some of those issues. We are also working on some of those issues at the IETF, the standards body for Internet protocols. It is all about trying to create an overall ecosystem in which standards, technologies and companies come together to actually take these research ideas and move them out into the real world.
Q11: I know that in the optical space in particular Intel made several investments and acquisitions. Can you talk about how that extends Intel's capabilities?
A11: We have acquired some companies that have expertise in optics. They bring an established product and a certain amount of research capability. We complement that with an investment in basic research, utilizing people and resources at Intel. One area in particular we're looking at is how to marry optical technologies to existing Intel competencies in high volume manufacturing. That is a large part of what we can bring to the game.
Q12: How would you describe Intel's approach to research?
A12: Intel is a very product-oriented company. That is our culture. We connect basic research that is done in universities and government labs with basic research that we do. We take that back into the industrial community to create the necessary ecosystems for the technology to flourish into real products. Ultimately, we contribute to the development of industry standards, and build products that deliver useful technology to the consumer.

For us, research is a critical piece in a chain that ultimately delivers value to consumers. Intel's take on this product-oriented philosophy is unique because we believe that the entire value chain is important. If you can't put that whole chain together, you may have an interesting lab, but you won't have true technology leadership.
Q13: What kinds of cool things can consumers look forward to in the future?
A13: I think our long-term goal is to make any intelligent product (and that's virtually any electronic product because they all have some sort of a processor) that has any reason to communicate able to communicate when it needs to at an appropriate data rate. This may be very, very low. I mean, your refrigerator probably does not need multi-megabytes per second. On the other hand, it may need a couple of kilobytes per second periodically to burst out some information or receive information about energy, either to shut down right now or try to limit its energy consumption because it is a peak period. That kind of communication does not take a lot of data, but it does require communications. So we're working on the notion that any electronic device that needs to communicate should be able to do it without the user having to get involved.
Q14: What about the communications devices people use?
A14: Today it is possible for most people, if they want, to stay in touch most of the time. They have a home phone, a cell phone, a pager. They can take their mobile computer and plug it into their phone at home, the Internet at work, or they can go to a coffee bar and plug in a wireless card and talk there at a hot spot. I keep saying "they can" because they have to do something to stay connected. That is a long way from having systems that are able to figure out for themselves the most effective way to communicate, configure themselves to do it, and then just do it.

Our goal is to make communications disappear as a consumer concern, in much the same way as your telephone system has disappeared. Most consumers use the telephone every day of the week and never think about how the phone system works. Data communications they have to think about very much. And we all know about cell phone hell. Fortunately, we can use the off switch. A nice thing about electronic mail is that it doesn't demand immediate attention. There are lots of communication modalities that I think that can be very compatible with our lifestyles, yet won't need to be constantly on call unless we want them to be.

As we continue to integrate technology that brings costs down and it becomes possible to put these capabilities into affordable devices, I think it is an achievable goal to have more and more consumer devices that will communicate in multiple ways, be easier to use and much less invasive.
Q15: Can you predict any particular products that we may see these kinds of communications capabilities in?
A15: I think it's basically the evolution of the products we have today. The PC is going to remain the center of most digital clusters and environments. It is a very powerful general purpose device that can lend a lot of intelligence to the things around it. I think you are going to see a whole raft of small devices for the home. We are already seeing some of them. Things like digital picture frames, MP3 players, the integration of your music system with telecommunications so you can download music and play it over your nice stereo system through your home speakers. Those things are certainly beginning to happen today in the enthusiast space. That kind of technology I think will become more common and available to people. Devices that let people have casual information. Maybe it's that digital picture frame that has rotating pictures of kids. On your way out to work in the morning, you tap the frame when you go by it and it gives you the road and weather conditions. It is communicating. It is convenient. It is a lifestyle improvement. So there are lots of those kinds of devices I think you will begin to see.
Q16: What are some of the biggest challenges you think Intel faces in terms of developing these products?
A16: I think challenges are what I alluded to earlier how to make the need for all these actions disappear. It's relatively easy to make stuff appear; it is much harder to make it disappear. Seamlessness, ubiquity, and getting costs down to where it makes sense to include the capability on a device. Today, it's a $100 investment to put an 802.11 wireless LAN card on a system. That's not trivial, nor is it unusually expensive. However, it is not the sort of thing you would include on a $150 audio device. So getting costs down is a big hurdle.

But probably the biggest part of the challenge is the overall integration of the components and technologies, making it simple and transparent for the consumer. One of the things we are talking about is the absolute explosion in the number of devices that have to communicate. Another aspect is the rather modest number of devices that communicate on the Internet today and yet the high management overhead associated with that, it would clearly be unconscionable not to improve the autoconfiguration capabilities.

The good news is that, if you go back to the early days of the telephone system, I think it was fairly convincingly proven that before long, nearly everyone would have been working as telephone operators. Luckily, technology came to the rescue, we automated all that stuff, and today there are just a few telephone operators, and they do add a valuable service. In a sense, this is analogous to what has to happen here. We essentially have to make almost all of this automatic.
Q17: I keep thinking about the VCR with the 12:00 o'clock flashing.
A17: That is a good example. How do you conceive of fixing that? The common wisdom is to make it simpler to program your VCR. A much simpler way to fix that problem is to not have to set the time. If it was a communicating device, it could ask somebody for the time. I mean, there are two approaches to that. One is that you can make the task easier. The other, and I think the preferable approach, is to make the task go away.
Q18: Can you try to put into perspective the activity that goes on in the communications and networking sphere and how this fosters Intel's technology? How does this fit with what Intel does and with Intel's vision of the future?
A18: Intel has adopted what we in the lab call a stack model of thinking. What we mean by that is that you have to recognize that technology never exists in a vacuum. It is not about just a piece of technology, it is about that technology appearing in the context of a whole set of necessary ingredients. At every level you ask if this interesting technology innovation is going to be valuable to somebody, and then you ask what other stack elements are going to have to fit with it.

So I think in terms of how we foster innovation, we are doing the necessary basic research to advance pieces of the technology that are our expertise. We are tying those technologies into the overall ecosystems and standards that are necessary to make a difference in the world. I think if you look classically at Intel you see our traditional strength as a PC building block supplier. These days, PCs must communicate. Communication technologies are critical to the advancement and harvesting of the full value of PCs by consumers. If they don't get value out of it, we're not going to be successful. Communications is a huge piece of that.
Q19: And what about the communications industry in general?
A19: That's definitely our other concern, communications as its own business. Over the past few years, Intel has become a major player in the business phase of communications products, and we are doing the research that fits into that. What we would like to be able to do is take the work that we are doing, put it in the context of the other necessary pieces, and then help the entire industry move forward.
Q20: Given the challenges of being in the communications business right now, why do you think Intel will be successful in this space?
A20: I think it depends who you compare us to. There are certainly a number of what I would call old line communications companies that have had a lot of difficulty with the transition to the Internet. The Internet has really changed the rules for a lot of things. Partly because of the way in which new technologies are involved, and partly in the way that it has changed who the players are that deliver communications services. The days are over when the only purchasers of communications equipment are a handful of very large telephone companies, all of whom are regulated, and know exactly who their consumers are.

The overall shift is to a much more open environment with many companies in the communications space, and a customer base that has changed a lot. Intel has a long history of adapting to change and serving chaotic markets, like the PC market. The Internet is the ultimate of chaotic technologies. If you look at the history of data communications over the last decade or so, the Internet ran up against Token Ring, a very structured technology. The Internet ran up against ATM, the telephone companies' answer to communications and again a very structured technology, and one that some would argue had attributes the Internet could not match. Nevertheless, the Internet won. I think we are going to see that type of trend going forward. It's interesting that people are able to take real advantage of chaotic technologies to do things far beyond what they originally thought about doing. I think Intel is uniquely positioned, because the PC business required exactly the same sort of thinking style.

Back to Top

** All information provided related to future Intel® products and plans is preliminary and subject to change at any time, without notice. No license, express or implied, by estoppel or otherwise, to any intellectual property rights is granted by this document. Except as provided in Intel's Terms and Conditions of Sale or License for such products, Intel assumes no liability whatsoever, and Intel disclaims any express or implied warranty, relating to sale and/or use of Intel® products including liability or warranties relating to fitness for a particular purpose, merchantability, or infringement of any patent, copyright or other intellectual property right.
*Legal Information  |  Updated Privacy Policy
©2002 Intel Corporation