Volume 11, Number 1, 2002

Where timing is everything

David Mills' friends and colleagues often refer to him as Father Time, but that's not because he sports a white beard. Nor is it because the professor of electrical and computer engineering has a "backwards" clock on his office wall that tells time in mirror image.

Instead, the reason for the nickname is that Mills is a member of what he calls the "Internet Construction Corps," a group of about 10 researchers who began working together 25 years ago to develop and implement Internet protocols, and his area of specialty is the Network Time Protocol (NTP).

"NTP is arguably the longest running, continuously operating, most ubiquitously available protocol in the Internet," Mills says.

The protocol enables precise time synchronization on the Internet. Without it, online experiments could not be precisely measured and controlled, stock market buy and sell orders could not be timed, and web streaming of video would be chaotic. NTP makes possible such online activities as aviation traffic control and monitoring, radio and TV programming launch and control, multimedia synchronization for real-time teleconferencing and traffic engineering. The time function on the Internet is as essential as the clock on your wall or the watch on your wrist, Mills notes, but it is far more complex to set and maintain. And, the consequences if it breaks down are far more wide-reaching.

For example, Mills says, as the federal government prepared to handle potential computer chaos when calendars rolled over to the year 2000, "I was summoned by the Office of the President to consult about the Y2K issues, of which NTP was a considered hazard. I met with the folks who operate the two dozen or so root servers. They advised me of the 'time-warp hazard' and considered it as a possible terrorist opportunity."

Although Y2K turned out not to be the major problem anticipated by many, keeping time on the Internet continues to be a critical issue, Mills says. However, he jokes that as his career in this field progresses, he knows more and more about less and less.

"When I started working on the development of the Internet," he says, "there was no base of knowledge in the field. All of us knew just about everything there was to know, and we kept up with it as it developed. Now, everything has become extremely specialized."

Mills remembers the early days of Internet development as resembling a big game. "The members of our research group were spread out all over the country writing programs," he says. "We'd talk to each other and try to work with each other's programs. We'd get a point if we could write a program that no one else could solve, a point if we could work someone else's program and an extra point if we could crash it. We sent out what we called 'Chernobyl packets'--programs that would expand as they were disseminated from one site to another--so that we could test the resilience of the system.

"The strangest thing about the whole process is that we were inventing e-mail, file transfer protocols and remote interactive access, using the very infrastructure that we were developing. In other words, we were building the infrastructure so that we could build the infrastructure. I learned the most important lesson of my life from this experience--that people who are actually going to use services should be the ones to build them."

It wasn't until 1986 that the Internet, initially funded largely by the Defense Research Projects Agency (DARPA), went beyond limited government and military use. In that year, the National Science Foundation selected several supercomputing sites throughout the country to bring the Internet to colleges and universities.

Mills' work on this project has led to his selection as a fellow by both of the major professional organizations representing the two facets of his field--the Institute of Electrical and Electronics Engineers (IEEE) and the Association for Computing Machinery (ACM). Mills explains that IEEE, of which he was named a fellow in January, reflects the engineering side of his work, while ACM, which named him a fellow in 1999, is the computer science and information technology side. He admits that it's very unusual for one individual to be recognized by both professional organizations, as most people select one path or the other.

"But, I chose to walk right down the middle," he says.

Mills continues to refine the critical tools that he has developed. He remains challenged by the desire to make timekeeping ever more precise and to improve its security.

"Having the wrong time could completely destroy the network," Mills says. "My job is to try to harden it against a cyber-terrorist attack."

What's next for this man who is a slave to the clock? "We've conquered the Earth," he says. "Now, we're going into space."

Mills isn't kidding. Under a research program funded by DARPA and NASA, a group of researchers is developing the Interplanetary Internet, with a near-term emphasis on Mars exploration and mission support.

In this new venture, Mills still is the guy with an eye on the clock. While the rest of us are focused on the Eastern, Central, Mountain and Pacific time zones, the vocabulary of Mills' business includes International Atomic Time, Terrestrial Dynamical Time and Universal Coordinated Time. And, while the rest of us are concerned about such things as the five-hour time difference between New York and London, Mills and his colleagues are addressing such issues as transmission delays between Earth and Mars, highly variable transmission speeds and the logistics of scheduling connectivity opportunities.

Mills also says he continues to have fun. "This reminds me of the early days when we were still conquering the Earth," he says. "I feel very lucky to be one of only about three people in the world being paid to do this."

--Diane Kukich, AS'73, '84M