By Rob Knies, Managing Editor, Microsoft Research
In Boston on April 17, in the Great Room of the Massachusetts State House, Victor Bahl, director of Microsoft Research Redmond’s Mobile Computing Research Center (MCRC), will be introduced as one of six recipients of the University of Massachusetts Amherst’s 2012 Distinguished Alumni Awards. The awards are presented annually to those who have built on their university experience to attain notable achievements in the business, public, or community-service realms. Bahl, 47, earned the recognition with a superlative career that has featured a series of firsts in networking research, and on March 1, 2011, he was named head of the newly formed MCRC. He recently took time from his busy schedule to reflect on his career and his professional motivations.
Q: How does it feel to receive an award like this at this stage in your career?
Bahl: It feels great! I’m honored and humbled by the award—and surprised.
This award holds particular significance. It’s one thing to become an IEEE Fellow or an ACM Fellow, because the path to those is clearer: You have to have a certain amount of technical impact, but we train for that. This one is special—it’s not about technical accomplishments only, and it comes from an entire university with so many different fields.
While it’s nice to get external accolades, though, it’s a lot more important to me that I do things that have a big positive impact on my field and the company, especially in areas that require deep and innovative thinking.
Q: What were your initial goals for the MCRC?
Bahl: Mobility and Windows Phone is an area that requires focused thinking about how we’re going to distinguish ourselves. The industry is big, things are evolving fast, and there is lots of money at stake.
When I thought about the mission for MCRC, the word “indispensable” came to mind. We had to invent technology that would become indispensable to people. I started thinking about fine-tuning some of the great ideas being pursued in our labs, getting them to a point where Windows Phone and Windows would productize them. The goal for MCRC researchers became “let’s not worry about publishing papers” but instead “let’s focus on working with product groups and creating indispensable technology for our users.”
The approach I took was to talk to our product groups to find out what problems they cared about most, then identify our existing research that addressed those problems, filling in the holes, and begin moving them into products. In parallel, we began pursuing a few blue-sky ideas that would move the dial in terms of how people think about Windows Phone and about Windows and Microsoft in general.
Q: Thirteen months after the debut of MCRC, how have you progressed?
Bahl: Quite well. We are in the process of moving five solid pieces of technologies into the next version of Windows Phone. By all counts, that is a major success. We’ve also initiated some bigger projects that, if they succeed, will generate lots of excitement and have a much wider impact.
I would give us high points in terms of succeeding in getting these technologies into the product.
Q: What is the technical vision that underpins the MCRC work?
Bahl: One part has to do with marrying the cloud and mobile. The mobile cloud is not a new thing; we started working on it many years ago. But today, the whole world talks about it.
People and companies think about the mobile cloud in two ways: They think about it as a storage place that enables you to access your data from any device. And they think of it as mega-services such as Bing or Exchange, which require thousands of people and millions of dollars to build and operate.
I believe the way things are going to evolve—and Microsoft is well positioned for this—is that there will be a library of cloud services that are available to developers, who will tie them together in creative ways to offer major new functionality to users.
Most people already understand what an app store is, so think of this as a “service store” for software developers. These services will be built from sophisticated computer-science algorithms that are both resource- and data-intensive. The services will run in the cloud for the benefit of mobile devices. A service-composition framework would make it quick and easy for developers to stitch these services together and build the next generation of amazing applications. You can be creative in how you use stitch these services together.
Microsoft has been a phenomenal company in helping developers, and this is a great development story for them. Now, you can build extremely powerful applications, using services such as optical character recognition, speech recognition, object recognition, trajectory prediction, big-data analytics, heavy-duty graphics rendering, and gesture recognition.
We’re also looking at how to partition computations dynamically between the cloud and the mobile devices. When the cloud is available, mobile applications should opportunistically use it for computation. Certain portions of the application will run on the mobile devices, but other computationally intensive parts will be run in the cloud. All this will happen seamlessly without the user knowing about it.
Effectively, we want to use the cloud to remove all resource constraints that prevent us from building super, amazing mobile apps that are able to enhance our cognitive abilities. We have a number of projects in that space, and we’re making pretty serious headway. More than 60 universities have been using Project Hawaii, and our plan is to make these services available to the worldwide developer community. Then, consumers will start to see the power we’re talking about.
A final part of the mobile-plus-cloud vision is that everyone on this planet will have Internet connectivity. We have been working diligently for many years to make inexpensive, ubiquitous Internet connectivity a reality. We have developed technologies such Wi-Fi hotspots, mesh networking, and, recently, white-space networking, all of which are aimed at bringing Internet connectivity to everyone.
Q: How did you end up at Microsoft?
Bahl: I joined Digital Equipment Corp. in 1988. It was a fantastic company. I really liked it, and I was lucky in that I got to work with some really smart people. Everybody around me had a Ph.D. at the time, but I didn’t. I’d always wanted a Ph.D., and being in a research group made it even more important.
Digital had a program that allowed me to pursue my Ph.D. full-time for two years with full tuition reimbursement, full salary, and perks. I was the only one selected for this program from a large organization, more than 120,000 employees at the time. It was a highly competitive and prestigious program.
After getting my Ph.D., I wanted to do more challenging work, and I was finding it hard to find that at Digital, because in ’97, the company was in uncertain times. So I started thinking about other places.
My brother was with Microsoft at the time. He said: “You should come and visit. At least, you can see your niece.” So I came and interviewed with Rich Draves and Rick Rashid, who at that time was heading the Operating Systems group. Rich and Rick told me that we were going to create a lab that would be better than Bell Labs, which was considered the best research lab in the world when I was growing up. Everything they said to me was just right, so I took a risk. I had several other offers at the time, but I decided to come here.
When I came, I was going to work in networking, and I told them, “Well, our group name has to change, because I’m doing networking; I don’t do operating systems per se.” I suggested, “Let’s call it Networking and Systems.” But there were five of them and there was one of me (laughter), so we ended up calling it Systems and Networking.
I found Microsoft Research to be an amazing place. Its managers had the confidence and foresight to let people be what they can be. I didn’t feel there were any barriers. I was only limited by my own imagination, my own hard work, and my own creativity. I received tremendous support and was able to build one of the best networking groups in the world.
Q: Let’s talk about some of your significant successes, such as making Wi-Fi publicly accessible.
Bahl: My first piece of work when I came to Microsoft Research was to bring wireless networking to Microsoft. When I got here, Wi-Fi was not popular. In fact, there was no Wi-Fi in Microsoft or in Microsoft Research.
My first objective was to deploy Wi-Fi throughout Microsoft Research. I did that. I was the first person to bring Wi-Fi into Microsoft Research, and it was so successful that Bill Gates decided we were going to invest in it, and the Microsoft IT department rolled it out to the rest of the company.
The second thing I did was talk to the Windows team about changing the programming abstractions for wireless cards. I explained to them that wireless is not like Ethernet networks, for multiple reasons, and hence must be treated differently. They were not convinced that the difference was large enough to change the abstraction and add new APIs, so I decided to build an indoor location system called RADAR. GPS doesn’t work indoors, and I wanted to show that if you abstract the wireless-network card out differently, you can get signal strengths, and when you get signal strengths, you can find patterns, and from the patterns, you can find where somebody’s located. RADAR became a groundbreaking technology.
My next big project was wireless hot spots. I used to argue that cellular bandwidth was going to become scarce so we needed an alternative wireless-connectivity solution. A majority of people spent most of their time indoors, so therefore, enabling wireless-LAN connectivity in public spaces was the way to go. But there were lots of skeptics, and people were gung ho for the emerging 3G networks. They thought that 3G networks were going to solve everybody’s connectivity problems, offering high-bandwidth data connectivity everywhere.
I was adamant that Wi-Fi would enable the same access, so I decided to build the first public Wi-Fi hot-spot network in the world. I deployed it in 1999 in Crossroads Shopping Center in Bellevue, Wash.
When you reflect back and look at the success of public Wi-Fi hotspots, it sounds crazy that anyone would argue against them, but they did. I realized that it was going to be big, and now they are everywhere—I feel pretty good about that.
Q: How about mesh networking?
Bahl: That was another interesting project for me. Even to this day, when you look at mobility, connectivity is the limiting factor—poor or no connectivity will prevent the massive adoption of mobile-computing technology. The telcos control this chokepoint, and they can limit the rate at which innovation happens. If you don’t have enough bandwidth, you’re not going to be able to access amazing 3-D images or watch movies from the cloud on your mobile devices
I remember vividly, the day Craig Mundie [Microsoft chief research and strategy officer] invited me to have dinner with him. I was surprised and curious, because he hadn’t told me what the dinner meeting was all about. Over dinner, he talked about his vision for creating another ecosystem for connecting mobile devices. From those discussions emerged our mesh-networking project.
Initially, I was hesitant, because a special case of mesh networks, multihop ad-hoc wireless networks, had been studied for almost two decades, but they were good only for vertical segments, such as for solders in battle theaters. But suddenly, we looked at them as an opportunity to provide low-cost Internet access in low-income-housing areas, in rural areas, or in developing nations, where it is not easy to build an expensive cellular infrastructure. That made it interesting to me, from a perspective that more and more people could get connected.
I spent three years building and popularizing mesh networks. Our group did so much amazing work there. We deployed meshes in local communities and produced software kits that we gave to over a thousand universities worldwide. A whole industry was spun off that.
As mesh networking became more and more popular, it became clear to me that it was time for us to step back. One of our jobs as researchers is to open up new fields. When others see the path and the end game, they become interested and begin working, and they are then sometimes in better position to help mature and build the technologies further.
Q: You’ve also been an advocate for unlicensed use of portions of the electromagnetic spectrum.
Bahl: Spectrum is a scarce and finite resource, and our capacity to consume it is going up incredibly all the time. More and more bits need to be accessed. The question is: What are we going to do about this? Since 2003, I had been involved in pushing the FCC to make more spectrum available.
We believe that unlicensed spectrum has created a tremendous amount of innovation. Wi-Fi is one example, Bluetooth is another, ZigBee, baby monitors, ham radios … All this operates over unlicensed spectrum. We were writing lots of papers on the topic and telling the FCC that having access to greater unlicensed spectrum is important for connectivity and innovation.
Then something interesting happened. The U.S. Congress’s mandate on transitioning analog TV to digital TV freed up some spectrum. Lots of telcos jumped in and asked the government to license that spectrum immediately, saying: “Hey, we’ll buy it from you. We have billions of dollars.” But we were saying, “No, no, open some of this spectrum for unlicensed use, because it fosters creativity that leads to many new forms of connectivity.”
Then came the powerful notion of opportunistic networks. The idea was that if the owner or the primary user of the spectrum is not using the spectrum, let someone else use it, and if the primary user comes back to use it, we’ll just move along and find some other open space. It just makes perfect sense; we do this in life all the time. This notion of primary users and secondary users started to emerge. We needed to prove that you could build real-world networks that use spectrum opportunistically. Going into this I had one goal: to get the FCC to change its policy to provide more unlicensed spectrum, and we accomplished this goal by building the first white-space network in the world. This was the first network that conclusively proved that it was possible to build real-world opportunistic networks.
We got an experimental license from the FCC, we built our network right here on campus, and then everybody wanted to see it. Spectrum regulators from all over the world visited us. Even the FCC chairman came to see our network. In August 2010, I spent a Saturday afternoon with him and Mundie and explained that this was going to be the new world. The old way of divvying up spectrum into chunks, we were saying, is not the way to do things. The way to do things is to let networks use spectrum opportunistically. If you do, you will get rid of the spectrum problem. We were basing this on an observation that others had made by looking at spectrum analyzers: Most of the time, frequencies actually are unused.
Apparently, we were convincing. Less than a month after meeting with us and seeing the network in operation, the FCC approved the use of white spaces. We had set a goal, we had met it, and we had helped to change U.S. policy.
Q: What is it that enables you to look at these difficult problems from a unique vantage point, one that enables you to take on these challenging problems and solve them?
Bahl: I travel a lot and meet lots of smart people. We have labs are all over the world, and I give a lot of keynote talks. In my travels, I talk to a lot of folks, and by being open-minded, I’m able to see problems.
Recently, I attended the Mobile World Congress, a large event that brings technical and business people from all over the world to one place. I attended several panel sessions led by mobile-app developers, and I could see a tremendous amount of pain in their eyes. We have gone from a world in which the web which was open to one that is proprietary. We’ve created silos: iPhone, Android, Windows Phone. Developers are forced to choose one of these silos to work in, and that’s painful, because they would rather write software once and have it work everywhere. That’s the world they want to be in. I have a deep urge to try to solve that.
Listening to lots of people and taking time to analyze what I have heard helps me formulate projects. The trick after that is to distill the vision to manageable components with a crisp problem statement that I and others can work on. I keep working on these problems until I have found the solutions. I have a strong conviction about taking things to completion.
Another way is to do this is bottom up. I work with smart researchers who work on cool problems. I often look for common threads between their projects, which helps me create a compelling vision.
We work in an amazing field, a field that allows us to be creative and to solve real problems. How cool is that?
Q: Finally, what’s the best advice you’ve received?
Bahl: A lot of smart people have talked about having passion, conviction, and confidence. This is all true.
To that, I will add determination and perseverance. If you honestly believe in your idea, go get it done. Teach yourself to be fearless, and do not rest until you have accomplished what you set out to do.
More advice: Every so often, take some time to think. In default mode, our lives are fast-paced. Try not to be in action mode all the time. Take time to educate yourself, reflect, and think deeply.
I once asked Jim Kajiya [a Microsoft distinguished engineer who works for Microsoft Research Redmond], “What advice would you give me?” He responded, “I don’t do things that I think others can do.”
As a research manager, as I think about projects and what to work on, I try to remove myself from the equation and see who the best person would be for the job. If I can’t find the right person, that’s where I go and help. I’m thinking about how to be successful in the project and then filling in the holes.
My dad was a great fan of Napoleon Bonaparte, and Napoleon Bonaparte had a famous saying, that the word “impossible” is in the dictionary of fools. The harder the problem is, the more interesting it becomes for me.