SM: When you left Foundry in 2004, what was the purpose of your departure?
LC: I stayed there for eight years, and I had never stayed anywhere that long. Once the company was well established, and I realized that I had worked in all of the various groups possible, I knew it was time for me to leave.
SM: Were you planning on starting your own company?
LC: I knew I wanted to eventually, but when I left Foundry I did not know what I was going to do in that regard. I knew it had to involve my engineering background. I also looked at what motivates me. First, I am competitive and want to build the very best product in the industry. The second aspect was realizing that when you build a company and take it to IPO you can really benefit a lot of people.
I was also looking for the aspect of a personal challenge. I was trying to find and create something that people said could not be done. Some people in the industry told me you could not start a company with one founder anymore. I did not think that was true and decided to do one on my own. I also wanted to make sure whatever I started had a global perspective.
SM: You eventually figured out what you were going to build. Can you talk about that process?
LC: I decided that we needed a better web application delivery service, especially in the wireless arena. We use way too much power today. Mobile devices have 1 billion connections to the Internet. Rich media is increasingly dominating the Internet, including delivery to mobile devices. I decided to build a product which would address the web application delivery market.
SM: In the 2004/2005 timeframe there were other players in the web application delivery arena. How did they factor in?
LC: I look at it like Google. They were not the first ones involved with meta search, but they are now the biggest. I realized we could compete not only in terms of product innovation, but also in terms of a business model. A lot of companies were using software modules which they would license. I know the analysts liked that, but customers told me they really disliked the licensing. It was a very easy business strategy to provide a license.
I went to some of the largest carriers in the world. They explained that their electricity use, user traffic, and storage requirements had all quadrupled in the past three years. That was acute movement. Customers were paying a fixed amount for bandwidth, yet their bandwidth requirements were getting higher and higher.
SM: So you focused on power as a key element?
LC: It was important. That gave us the opportunity to build a green appliance. Our design, from the very basic elements, are green.
SM: What did you do with your architecture to enable you to offer power savings and a green device?
LC: There are a couple of things. First, if you provide the best performance then you can consolidate three or four devices into one. That saves power.
SM: What is the functionality of the device?
LC: We are expanding the traditional load balancing market into the web application delivery market. Our product does compete with load balancers but it does not have 100% overlap with that space. Our product is a web application delivery platform.
We started by building a very powerful platform. In 2005 the industry was still using single CPUs. Our focus has been on a multi-core, multi-processor, multi-CPU architecture. It can scale all the way up to 30 CPUs. We use Intel processors but are not tied to them. Our design is processor independent, which allows us to select a processor for a device based on best performance per watt.
SM: A multi-processor, multi-core system will definitely give you more performance, but if the chips are coming from someone who services multiple vendors, what is your defensibility? Couldn’t somebody else build those types of products using the same chips?
LC: The question will always come down to secret sauce. We have something we call Advanced Core Operating System. It took us two and a half years to build a kernel which was optimized for the multi-core, multi-CPU architecture in the web application delivery market space. We can easily distribute the traffic regardless of the traffic type.
This segment is part 5 in the series : Still Innovating in Networking: A10 CEO Lee Chen
1 2 3 4 5 6 7