Ada Lovelace Day Spotlight: Radia Perlman

The next time you visit Facebook, do a search on Google, waste time on YouTube, or do anything online, stop and thank Radia Perlman.  It’s her algorithm that makes it possible for data to traverse the Internet, thus earning her the nickname “Mother of the Internet”.

Dr. Perlman is a Fellow at Intel, and has spent her career working on computer networking and computer security.  She earned bachelors and masters degrees in mathematics from MIT in the 1970s and a PhD in computer science from MIT in 1988.  She’s worked for many big-name companies throughout her career, from BBN to DEC to Novell to Sun.  She has nearly 100 patents in networking and security technology and has authored two widely-used and innovative textbooks.

(On a personal note, I saw Dr. Perlman speak at Grace Hopper in San Diego in 2006—she presented some of her more recent security work.  The talk was very interesting and thought-provoking, and afterwards I got to speak to her for a few minutes one on one—one of the highlights of that conference for me!)

Dr. Perlman’s most famous contribution—the one that earned her her nickname—solved a particularly thorny problem in the early days of computer networks:  how do we ensure that data gets to its intended destination without getting hopelessly lost, or going in circles?  How can we determine the path that data should take when going from one place to another?  (If you consider for a second just how many millions of computers, smart phones, appliances, and other devices exchange information over computer networks—well, you get a sense for just how important this problem is not only to solve, but to get right!)

Unfortunately, her first attempts met with….well, I’ll let this snippet from a 2006 article in Investor’s Business Daily tell the story:

Radia Perlman had a solution for an information routing problem. Unfortunately, no one was listening.

It was the mid-1970s, and Perlman was a software designer for computer network communication systems — and one of the few women in the field.

At a vendor meeting where engineers were asked to help with the routing problem, Perlman spent 30 minutes illustrating her solution with an overhead projector. But the event organizers ignored her findings. Why? Because she was a woman. What did women know about computers?

“At the end of the meeting, the organizers still called for a solution after I had just given them one, which really irked me,” she said.

She persisted, though, because she believed in herself and she also believed in her solution.  Eventually she made her way to DEC, where her solution met with more receptive ears….and the rest is history.

In a nutshell, her solution is what we call a spanning tree.  Let’s say that you have a bunch of computers that are connected together, and that you want to send data from Computer A to Computer B.  There are most likely multiple routes the data could take, because there’s more than one way to get from A to B.  Spanning tree ensures two things:  (1) that there is a way for data to get from Computer A to every other computer in the network in a “straight line” (i.e., without doubling back to a computer it’s already visited on its journey), and (2) that these data “paths” are the shortest and/or most efficient ones possible in this network.  So not only does spanning tree ensure that my data can get from A to B, but that it will do so in the most efficient way possible.

The really cool thing about this algorithm is that the computers in the network figure this all out themselves, without human intervention.  This means that if something happens to the network—a computer crashes—the network can figure out a new spanning tree all on its own, as soon as it figures out that there’s a problem.  This sort of self-healing behavior is what makes our modern-day, large, sprawling Internet possible—without it, things would surely ground to a halt often, because computers are always failing and crashing and otherwise misbehaving.

Dr. Perlman has made many other fascinating contributions throughout her career—for instance, I just learned that as an undergraduate, she worked on ways to teach computing to very young kids using LOGO and TurtleGraphics.  Her more recent work is more heavily security focused, studying new ways to encrypt and decrypt data, make files “disappear reliably”, and do distributed authorization. In 2005, she won a Woman of Vision Award for Innovation from the Anita Borg Institute for Women and Technology.

Throughout her career, her work has been ahead of its time, and I look forward to continuing to follow her career and her contributions.

(Post for Ada Lovelace Day—thanks Suw Charman-Anderson for organizing it again this year!)