You are here

Internet's gatekeeper

Akamai Technologies CEO Tom Leighton talks about how his company keeps online traffic fast and secure.

BT_20160924_ACAKAMAI24CL4S_2505116.jpg
'You (the CEO) have a debate, ask for all the ideas and poke at them and if there is any disagreement you have to make a decision and then you have to move forward as a team. Sometimes you make mistakes but those get fixed.'

IT'S not often that a group of academics, enamoured of interesting but obscure (to the layman) mathematical theorems and the popular TV franchise Star Trek, end up setting up one of the most important technology companies that you haven't heard about - a multi-billion dollar company responsible for distributing almost 30 per cent of all the traffic on the Internet.

Akamai Technologies, which runs a content delivery network over the cloud, operates a network of servers around the world. It rents out capacity on these servers to customers who want their websites to work faster by distributing content from locations close to the user. It also provides security by preventing one of the most common methods of attack on websites, known as DDoS (distributed denial of service). DDoS is a type of attack where thousands of compromised computers are used to target a website or server at the same time, causing it to overload and become unavailable.

Tom Leighton, Akamai's CEO and co-founder, says the company develops software and places it "everywhere we can" on the Internet.

"So it's on a couple of hundred thousands of our servers which are at thousands of locations on the Internet and it's also on hundreds of millions of devices in homes and offices around the world.

Data protection

"Whenever any one of the two billion people with access to the Internet go to websites, they are getting our software being put on their devices for the time being. Billions of devices have our software running on them... Our software is set up to make communication very fast and reliable. So if you click on something, we'll make it show up on the device much faster. If you are watching a video we will make the quality better. We also protect the data and the communication."

Most big banks, major sporting events such as the Olympics, media outlets and government agencies are customers of Akamai as they face the maximum threat from DDoS attacks.

Explaining why these entities turn to Akamai, Mr Leighton notes that there is a whole spectrum of security issues. "We develop a lot of the defences. Take the Olympics: if we wouldn't have been providing security they would have been DDosed (resulting in the website crashing or being taken over) because many entities tried to do that. (Because of us) you could watch the Olympics and you got the real content and not somebody else's content."

He says that the DDos is the most basic type of attack, where "somebody tries to deny access to the site".

He adds: "Today we are pretty unique in being able to defend a site against the biggest attacks. Banks are targets, governments are targets, as are big media events and they all can stay online today because of us."

At the next level there's the application layer attack, where somebody is trying to come into the website, maybe a government site or newspaper, and put their story all over the site, in effect defacing the site, he says.

"We stop that from happening and at the same time you have attackers trying to come in and steal confidential information from the side, maybe credit card details. We also stop that very successfully today."

Now, adds Mr Leighton, where Akamai doesn't yet operate but is about to, is on the enterprise side. "There's phishing, malware, data exfiltration from an enterprise. There the bad guys are well ahead and you can see it all the time. I had all my tax history stolen. And you know, I think many Americans have had their whole medical history stolen. The (US) government lost all records about everybody with a security clearance."

All this is outrageous, adds Mr Leighton. "We are now developing products, it will take us years to be able to stop (the bad guys) but we are starting with our first line of products by the end of the year. This will start raising the defensive barrier."

Talking from the perspective of users, he notes that you just can't add software (like anti-virus, firewalls and other security software) on your site and be done with it. "That's the traditional approach and it doesn't work anymore. There is no way you as a company could afford to have all the connectivity that you would need to prevent a DDoS attack (which could involve several hundred thousand computers accessing the site at once).

"That's why Akamai is unique in being able to stop large attacks because we intercept the attack traffic where it starts everywhere around the world before it ever gets to the data centre or server hosting a particular website.

"There's a fundamental change that's happening to security. You can't just defend your data in the data centre. The defence has to take place in the cloud but not just any cloud; a giant platform like Akamai's has to be able to defend you. So companies hire us and it's the service we provide that all the traffic that is directed at their site comes to us first and we filter out all the DDoS attacks."

He notes that the company acts like a sort of gatekeeper on the Internet. "All the inbound traffic comes to us and companies also hire us to deliver the content over the same platform.

"When you go to any major website today, say a major media organisation, your request for a particular piece of content through your browser comes to one of our servers that is near you. The first time there's a request for content, we go and get it from the (client's) data centre. But after that we spread it wherever it's needed and hand it out to whoever is asking. We have enormous capacity to do that and we do trillions of deliveries a day. It's the largest scale thing out there.

"And that same platform intercepts all the attacks as well, so nobody gets to the origin of that site. All the interaction is with us and we have enormous scale because our software and servers are at the edges of the Internet where all the capacity is."

Akamai has around 6,000 employees worldwide, is present in 65 cities around the world, and has a run rate of around US$2.5 billion (US$2.2 billion in 2015). According to Mr Leighton, the company is very profitable with a 40 per cent Ebitda (earnings before interest, tax, depreciation and amortisation) and a high gross margin in the region of 70 per cent.

The company has come a long way from its accidental birth at Massachusetts Institute of Technology. It started out as a research project at the university. In 1995, Tim Berners-Lee [LINK]- the person credited with inventing the World Wide Web (WWW) - had his office down the hall and he was worried about the flash crowd or hotspot problem, Mr Leighton recalls. A flash crowd is a large spike or surge in traffic to a particular website. Mr Leighton, who was then head of the Algorithms Group at MIT's Laboratory for Computer Science, was intrigued by the problem and believed his group could develop a solution.

"That was before there were flash crowds because in 1995 there was very little use of the Web. And my group was interested in routing problems for very large scale graphs from the maths perspective. We didn't write code, we proved theorems about algorithms.

Making history

"There was definitely a relationship between what would happen in a flash crowd and how you would be able to route big information to a network to avoid congestion. So it seemed like a very interesting problem for us to work on."

There was also another reason why Mr Leighton's team was interested. "It was an applied kind of problem. Since we were theoretical computer scientists, that was very helpful for us to get grant funding. Besides it would be potentially very useful and so it was very exciting for us to start working on that.

"The rest is history as they would say," he adds.

Mr Leighton gives an interesting perspective of how technology was viewed in the rarified arena of academia back then.

"I remember thinking as a graduate student at MIT, in the late 1970s and early 1980s, why doesn't everybody have email? We were on the DarpaNet - one of the few nodes - and at MIT we had electronic games then and that's the way the world was for us and the question was: Why wasn't it more ubiquitous?" DarpaNet is a term sometimes used for the ARPANET, the early network from which today's Internet evolved.

"I think that you could predict back then that you would have this kind of power in the devices (such as today's smartphones), there was Moore's law which predicted where chipset power would logically go; we already had calculators back then.

"The communication bandwidth, it was also logical that it would continue to improve. Even the notion of cellular devices, I remember being at a Darpa (Defence Advanced Research Project Agency) meeting way back when the then head of Darpa - there was Star Trek back then where there was a communicator - he said we should all have these, there is no reason we should not have it."

He adds: "Being in MIT, even in the late 1970s and early 1980s, yes we anticipated all this is going to happen. What was hard to foresee are the applications like Facebook, Twitter and Uber - who could have imagined those? But the actual infrastructure and the devices being there, they were a natural projection."

Mr Leighton provides this context to explain how, in 1995, bottlenecks were anticipated even before the Internet became mainstream and why Akamai developed a solution to take care of this particular problem. He recalls lamenting back then that it would take 20 years - "the typical technology time" - for an idea which was understood in the "ivory tower bubble of academia" to develop mass appeal.

However, the dotcom bubble happened and things got speeded up due to a lot of investment money becoming available. "I remember we (he and Daniel M Lewin, then a graduate student at MIT who worked with Mr Leighton) didn't want to start a company and thought that our technology could be useful and being mathematicians and theoreticians, that was a rare event. Usually we do things that are interesting but wouldn't be useful anytime soon.

"We actually tried to get others, like carriers, to use the technology first by licensing it and ultimately giving it away to them. Their reaction was that distributed computing doesn't work in the real world - everybody knows that. They literally used the phrase: 'Please go back to your ivory tower and we will move on'!"

Mr Leighton had to make a decision. "Do we return and continue in academics or should we take the plunge, leave MIT and start a company?

"The timing was very good because there was money available to start a company and a lot of interest in the Internet then. We took the plunge and launched the company in 1998. And that probably took many years off the technology transfer period."

He adds: "Without the dotcom bubble we might not have started a company as we were reluctant entrepreneurs."

The next year Akamai launched its first commercial service and went public with an IPO (initial public offer) in October. In 2000 Apple and Akamai announced they were combining technologies to build the backbone for Apple's Quicktime TV Network.

While things were progressing, the dotcom bubble burst. "It was extremely painful, we learnt how to have a budget and that we needed to be profitable - in hindsight it probably did help us to become a real company. It nearly killed us but there was value in that. There was very little funding for innovation during that period."

September 11, 2001, nearly broke the company. Mr Leighton's co-founder was the first person who died in the tragic 9/11 plane crashes into the World Trade Center in New York. "He was on the American Airlines flight 11 and he was killed trying to defend the cockpit and a stewardess, the first casualty before the plane crashed," remembers Mr Leighton.

"Danny was our spiritual leader and an unbelievable human being, it was brutal beyond belief," he recalls.

However, that day was also the busiest day for the nascent company, as it had to support the unprecedented traffic to various news sites from people looking for information about the tragedy. "So many news and government sites were getting crashed with either traffic or attacks and they turned to us to get them back online which we had to do. So everybody was working even harder than normal to take on that task, even in the midst of our personal grief," the Akamai chief says.

The two years following 2001 were "extremely difficult for the company", say Mr Leighton. "We were doing layoffs; we laid off two-thirds of the company. We got down to a few months to get by in terms of cash. It was a very hard period for us and still defines us in our character today. We are very frugal with money, we don't waste it and we have a very strong survival instinct. We are very tenacious and that helped us get through that period as well. But it was incredibly hard."

Akamai broke even in terms of cash in 2003 and became a "little bit profitable" in 2004 and "at that point we could see a clear runway". Mr Leighton adds: "Internally I always believed we had a path through it, even though everybody thought we were dead but we just didn't know it.

"That's part of it: we could see a clear path through and it was logical; that helped us internally to rally around that. We followed that path and then we grew steadily from 2004 to where we are today."

Mr Leighton was Akamai's chief scientist for 14 years before stepping into the CEO's shoes in 2012. The previous CEO decided to retire and the company did an executive search for about a year and during the process Mr Leighton realised that he wanted the job and so raised his hand.

As for the difference between academia and a CEO's role, Mr Leighton says that the culture in the company is very academic in the sense that employees do a lot of questioning. "The questioning is really vital. Open discussion and airing of ideas transparently is great. What matters is not your rank but the quality of your idea.

"At the end of the day (the CEO) has got to make a decision. You have a debate, ask for all the ideas and poke at them and then you make a decision and if there is any disagreement then the person with rank has to make the decision and then you have to move forward as a team. Sometimes you make mistakes but those get fixed.

"I think the openness, the questioning, the honest discussion is important; it makes a huge difference in making good decisions."

Mr Leighton says that Akamai is working on four things that it internally calls grand challenges.

The first is enabling a network where billions of people can, in prime time, watch a high quality video over the Internet. "That's a factor of 1,000 more traffic (volume) than that exists today," he notes. The second is to ensure the devices on a cellular network in a crowded city, "where you barely have the bandwidth to make a phone call", would be able to instantly display a webpage when a user clicks on it. He notes that even 5G will not be able to provide that kind of bandwidth for every person in a crowded city; the infrastructure will have to be improved.

The third grand challenge is security. "How do you keep your data safe across the board? How do you stop the cyber attacks? That's really hard but I think we can do that and at some point in the future it will be a safe world."

The final area is enterprise networks. "Today they are very challenged. The enterprise network manager's job is completely changed. The applications have moved to the cloud where he or she doesn't control them. The network itself is very expensive and has limited bandwidth and is performing very poorly because it's congested, because people in the branch office and employees on the move want video, need to access the Internet for all the applications to do their jobs. They need a factor of 100 more capacity. You can't afford that. So how do you enable that for the enterprise and be secure? I think we can help do that," he adds.

Mr Leighton still takes some classes at MIT because he thinks that it's important to keep in touch with academia "and it's important to bring industry into the classroom and teach the undergrads what they need to know to go out and create the next Akamai or work in established companies to make the Internet better".

Star Trek's special role

Not surprisingly he is a science fiction fan because "what's today's fiction is often tomorrow's reality". He's a great fan of the Star Trek franchise and still loves watching the shows. He shares an interesting anecdote: "We got a lot of excitement among the undergrads who did the first prototype at Akamai. We contacted Paramount Digital Entertainment in 1998 and they had the Internet rights to the (Star Trek) TV shows. They said if we built a service for them they would use it to put the TV series online and we would deliver it.

"They sent us two gigantic boxes of Star Trek paraphernalia. We distributed all those to the MIT undergrads and they went nuts and developed the prototype in a record time of a few weeks and that launched Akamai.

"So Star Trek has a special role in Akamai history," he adds. He firmly believes in the Star Trek motto: "To boldly go where no man has gone before."


Tom Leighton

Chief executive officer, Akamai Technologies

1957 Born in Arlington, Virginia, USA

1974-1978 Bachelor of Science in electrical engineering and computer science, Princeton University

1978-1981: PhD in applied mathematics, Massachusetts Institute of Technology (MIT)

Since 1989: Professor of applied mathematics, MIT

1998-2011: Akamai co-founder and chief scientist

Since 2012: CEO, Akamai