The mathematician who made the net faster

Akamai Technologies makes websites and online operations of companies faster and more efficient. Tom Leighton, professor of mathematics at the Massachusetts Institute of Technology (MIT), set up the company with his student Daniel Lewin, in 1998. Leighton, now CEO, recalls the backdrop in which the content delivery network company was born, and his charismatic co-founder Lewin, who died in one of the 9/11 plane crashes.

We started working on content delivery network technology without any thought of creating a company. In 1995, I was in the applied mathematics department of the MIT, my expertise being algorithms and theoretical computer science. But I would write papers, deliver talks, and supervise theses for students, not write code. Tim Berners-Lee’s office was down the hall from me. He was an insightful fellow and was worried about the architecture of the World Wide Web that he had created in the early ’90s.

Berners-Lee knew my group had worked on [the mathematical aspects of] large-scale networks, so he set us a problem to solve: If a website was on an origin server (the topmost in the hierarchy) and its content became popular, a surge in traffic, known as a ‘flash crowd’, could bring the server down.

I took up the project because it was challenging and would also help us get money to fund the team. If you’re a mathematician, funding research is a concern in the U.S. because math is usually at the bottom of the pile when they give out grant money. Funding goes into applied research and products because they bring faster returns. In mathematics, the chances of research being useful immediately are almost nil, because the journey from conception to technology can take up to 20 years.

I got more than $1 million (then Rs 3.5 crore) between 1995 and 1998 to hire students and pay them to do research. Danny (Daniel Lewin) came to MIT in 1996 and did some great work on algorithms to get around congestion, distribute content from a network of servers, and storage. All of it was mathematics; there was no coding. There was no plan to use it in practice, but that it sounded useful was good to keep the funds flowing.

In the fall of 1997, Danny’s neighbour told him about a business plan competition at MIT called the 50K Contest, which resonated with him because he was almost broke from his student loans and his kids were in private school. In those days, people in our line of research didn’t get paid well, and Danny (in his twenties) wanted to be a professor. He talked me into entering the contest. In an early round, we had to give a three-page description of the business idea. We won in the software category, and spent the $100 prize money on beer.

Preetish Nijhawan from the MIT Sloan School of Management was thrilled that we had won in the software category; Danny and I had no appreciation of that. Preetish talked us into going for the next round: writing a 30-page business plan.

I started reading up on venture capital from books borrowed from a library. Meanwhile, we started working with companies with whom we could develop the proof of concept. Danny also had been working on a “distributed computing platform” technology under which we would deploy our servers inside the clients’ networks, bringing their content and applications closer to end-users.

In the process of learning about the business side of our plan, we spoke to Internet service providers (ISPs) as potential customers of our technology. We wanted to make them more efficient, and they could save money while improving performance for users. But they showed no interest because they saw us as a group of mathematicians with a business plan; [instead of discouraging us,] that increased our conviction in the technology.

Finally, we were placed sixth in the 50K Contest. Later, we started talking to website managers because we realised our pitch wasn’t working with ISPs. Some of the website managers showed interest because our technology would make their sites faster. So, we created Akamai.

When Danny died in 9/11, I was devastated. He was my best friend, the heart and soul of the company. He had the rare combination of brilliance, drive, charisma, and the ability to lead and kick butt. Equally, his loss galvanised all of us at Akamai into coming together. Soon after, the dotcom bubble burst and people thought we weren’t going to make it as a company. We had to knuckle down and persevere and our technology was such that we were able to survive. We reduced headcount from 1,400 to 500.

It’s still hard to recover from Danny’s loss. He had presence and became a role model. Akamai’s biggest award is named after him.

We touched a billion dollars in revenue in 2010. So much is changing that there’s hardly time to breathe. Users want mobile devices to be as fast as the desktop. But the infrastructure is such that mobile performance is bad. Our goal is to grow to $5 billion in revenue by the end of this decade.

We also branched out into web security in 2011. When we listed in 1999, we disclosed that we would branch out into applications, security, and enterprise networks beyond the Internet. We have built partnerships with IBM and Riverbed, where our technologies are going into their enterprise local area networks. A lot more of that will be happening. This will become important with enterprise networks moving to cloud computing, and people bringing their own devices to work and accessing enterprise apps. The Internet is now merging with corporate networks to create a hybrid cloud.

(As told to Kunal N. Talgeri)