Q&A with Rock Star and Intel Fellow Ajay Bhatt

posted by Charlene Zvolanek on September 23, 2009

[1:27pm] Nervously awaiting the appearance of this Intel rock star. Last chance to submit your questions for the live-blogged Q&A with Intel Fellow Ajay Bhatt.

[2:05pm] Why didn’t you play yourself in the ad? - Anil Das

Acting is not my forte. I’m not really comfortable with acting. I do a lot of things well, acting is not one of them. Customarily, when you make a series of ads, it is good to use actors. What we were trying to portray is that we have some cool guys who work on some exciting stuff. Our rock stars are key technologists that are just as good in their own field. But they’re different from a regular rock star, and that was the concept. I was on board from the beginning.

[2:07pm] Prior to the ads did you find the same enthusiasm among the tech audience?

Among the technologists I’m very well known because I’ve been doing this for quite some time. Mind you, USB was originally conceived in 1992-93, so this is a little belated. I’m known for other technologies. I moved on years ago. USB goes on, but I moved up the chain. My role has been different. Again, USB is something that all users know. I use it. Everybody uses it.That’s why I think they picked USB as a case because the common person can relate to it.

[2:09pm] Are you your own user? Are you carrying anything USB today?

All the time. You’re never away from USB. My camera, my iPod, my computer, my storage stick. USB is everywhere. And by the way, originally I conceived USB as a user because I was trying to install something in the platform. And I struggled as a user. I said there has got to be a better way. That was my inspiration for creating. I said why can’t I make something that’s as easy as plugging in a table lamp and turning it on. Why does anything have to be so complex?

[2:12pm] Take a look at the wikipedia page for USB, and you’ll see no less than 6 connector types for USB (type A, type B, mini-A, mini-B, micro-AB and micro-B). I realize that “universal” refers more to the signaling schema, but how many more connectors will it take before the “U” in USB becomes a joke? -Steve Pomerantz

The name is somewhat dated. When we created USB, there was only one type. As the technology became very popular, more and more users came to USB. Some things that didn’t even exist when we invented it have come to reality now. And those requirements were not conceived originally. Clearly this is an evolution. Sometimes that evolution in a different market segment requires market-specific thinking. One thing is very clear, that when it comes to computers, most have A type connectors. The B type only refers to the device. I realize that forces users to buy a new cable, so I sympathize with them, but clearly it is an artifact of evolution.

[2:16pm] Every project, no matter how well-designed, seems to involve inelegancies that frustrate future engineers. Comment on how these things happen, and how they can be recognized and mitigated earlier in the process. -Paul Showwalter

There is no such thing as a perfect solution. When you define anything, there is always a compromise because you have to balance cost, complexity, usages, market needs. So I wish I could do some ideal thing that was perfect. Engineering is all about compromises. When you try to buy a house, there is no such thing as a perfect house. Even when you build it, there are always things that you would do differently as you learn about things after the fact. You make your best judgment based upon the environment you are in. I never thought USB would have this kind of longevity. I never thought USB would have such popularity. I was totally surprised by that. I knew we were solving an important problem. I never knew it would be so pervasive and so universal. And I’m saying this 17 years later. It’s been a while. It’s been a learning process.

[2:20pm] You’ve written about the hurdles in convincing senior decision makers at Intel to buy in on the work that the PCI Express Team was doing. At what point did you recognize the need for transition from AGP (Accelerated Graphics Port) to PCI Express? And how long did it take to realize that vision? - Anonymous

Right after we finished AGP I knew that AGP that had a limited life. Usually these things last anywhere from 5-10 years, and AGB was getting to the point where it was showing signs of age. So we knew we had to do something different. Once the opportunity was realized by us, then I wanted to subsume many more usages. We wanted to reuse the same technology for multiple users. It took me about a year to look at candidate technologies. It also took me a while to get to a point that I knew exactly what it was I didn’t want to do. I had to take time to determine what it was I wanted to avoid. Once that became clear, we started in earnest in 2000. By the end of 2002 we had finished the spec, and we had consensus with other people. By 2003 we had some early products. In 2004, we went into mass production. The rest is history.

[2:24pm] Would you consider that an average cycle from conception to market?

Usually conception to market takes anywhere from 2-3 years.

[2:25pm] Across any of the technologies you’ve worked on, are there key elements that you would do differently?

You will always want to change things. For example, USB—if I was to do this all over again I would make it a lot more power-efficient. But hindsight is 20/20. For AGP there are things we had to do for backwards compatibility. I wish I would have avoided them. There are deep technical things that I don’t have time to get into, like address mapping scheme; I would avoid that. On PCI Express we could have further cost-reduced it. So you can always do things better next time. You apply your best judgment, crisply define the problem, and go after that. Technology before its time is no good. Technology after its time is no good. So you really have to time it. There is a perfect window of opportunity that one must hit. You can’t be on the drawing board tinkering forever.

[2:28pm] Would USB ever be able to replace iSCSI/SATA/Fiber, particularly for storage, and create a true uniform interface for all devices? -Gagan Bajpai

The short answer is perhaps. Perhaps. There are no fundamental barriers, technically speaking, that would prevent one from creating a solution like that. The barriers may be related to business consideration, manufacturing, and managing the transition. I see most of the issues being business-related. Technology-wise it is pretty good on certain things, I suppose SATA and iSCSI might be much more suitable because they are specifically designed for storage. USB on the other hand is primarily designed for add-in devices from outside the computer to the computer. But there is reasonable overlap where I could see someone successfully doing something like that.

[2:31pm] Is there anything at IDF or anywhere in technology that excites you today?

A few things come to mind. Clearly the USB 3.0 camera down on the demo floor. It’s terrific. The quality; the speed. I’m quite excited about the developments on USB 3.0. I’ve seen some demos of wireless band personal area networks where you can create adhoc networks with Wi-Fi enabled devices that do the file transfers. And if you have support in your mp3 player, you could wirelessly sync your mp3 music list. Those technologies are quite intriguing.

[2:33pm] What was the most unexpected device that applied USB?

Oh there’s a whole host of them. There’s a whole host of devices that we never… USB charging was never conceived, but it’s so ubiquitous; it’s everywhere. That was pretty interesting. Even Apple adopting USB was quite a surprise to me. Apple was on 1394 Firewire, and they were so strongly behind that technology, but in the end when USB was done, Apple was the first one to deploy USB ahead of a lot of others. So that was a rather pleasant surprise.

[2:35pm] Why hasn’t Firewire taken off quite as much as USB?

Firewire is an elegant technology, but what we managed to do with USB was to manage cost, complexity, and feature set. So we had incredible focus on making sure it didn’t cost as much. On the other hand we didn’t want to compromise functionality or performance. First and foremost our goal was to make USB so inexpensive that we could give it away on every laptop. And we worked really, really hard with all the peripheral manufacturers. We gave them technical know-how and support to make them successful. So the grass roots efforts we put in was very important to USB’s success.

[2:37pm] It seems that Intel is good at partnering with other companies.

I personally believe that when it comes to creating standards we must be good citizens. We must bring our best ideas, and put them on the table, and make any standard that we participate in as good as possible, or even better. So we work really hard to help everybody out. Then when it comes to competition, we also feel that we are as fierce a competitor as anybody else. So we create the standards first, and once the standards are established we strive to make the best products that people would want to buy.

[2:39pm] How do you foster innovation on your team?

You foster innovation by being easily approachable. You constantly listen to people who have good ideas. At the same time you also listen to feedback from people who act as devil’s advocate. You listen to both sides of the argument. You propose your idea as a starting point. You always be open to new creative thinking that could improve your idea. If somebody has negative input you want to get them as early in the program as possible so that you can make course correction. Innovation is all about starting with a vision. Once your vision starts to materialize, and you turn it into the product, along the way you have technology reality. You need to think about who’s going to implement this. Who’s going to pay for it. Who’s going to use it. How much money are you going to make. How much money others are going to make. Will it make a huge difference to somebody’s life? So you really have to look at every aspect of your innovations, and then pursue it. It’s not just about technology. It’s also about technology in business realities.

[2:41pm] What’s next for you?

I am lately working on a new laptop. I would love to have a laptop that is totally untethered. That is totally wireless; even charging of the battery. Which is always on. Always connected via any means. It is always accessible. As soon as I open the lid its available for me to use and I don’t have to wait for my computer to respond to me. That’s my ultimate vision. I don’t know how long it’ll take, but I intend to keep working on it.

[2:43pm] Do you have any advice for the young engineers and developers out there?

I sure do. I would suggest a few things. Think differently. Before you start building something, decide what is it you don’t want to do. What do you want to avoid? Be very persistent. At the same time, be very open to feedback. Invite ideas from others, both positive and negative. And if the idea is good, accept it readily. If the idea is irrelevant evaluate it carefully and reject it. And next time the same suggestion comes in you don’t have waste time. Try to be open to others ideas and criticism because it will only make your ideas better.

Comments (0)

tagged: Accelerated Graphics Port, AGP, Ajay Bhatt, camera, Firewire, innovation, Intel Fellow, iPod, iSCSI, PCI Express, Q&A, Rock Star, SATA, standards, USB, USB 3.0, wireless laptop