Photo: Amr Awadallah. Used by permission.
Meet Amr Awadallah, one of four all-star co-founders of Cloudera.Although Cloudera is only three years old, it has the earmarks of a billion-dollar winner. The company just raised $40 million last November, bringing its total to $75 million and among its big-named investors is the Central Intelligence Agency’s VC arm, In-Q-Tel. The CIA is also a customer.
“We are building this to be a huge, huge company,” he explains. “My preference is to go public but it could be a big suitor taking it.” Indeed, rumours are circulating that Dell might be that suitor.
Cloudera customers are building applications that literally change what is possible with the Internet, based on a technology called Hadoop that was invented by Google and perfected by Yahoo. Hadoop can scoop up nearly unlimited volumes of data and analyse them, using low-cost servers and storage systems.
“It can do almost anything, with the sweet spot being data,” explains Awadallah. He explains it as “a data operating system that lets you join a lot of computer servers together and make them look like one big storage/compute resource.”
For instance, eBay uses Hadoop to improve search ranking. But it also powers Skybox Imaging, a company that takes massive streaming images from satellites and creates custom, realtime reports for customers such as how many cars are parked in certain lots or how many ships are in port.
Awadallah knows Hadoop like the back of his hand: he is the former VP of engineering for product Intelligence for Yahoo. He spent eight years building Hadoop systems before he left to start Cloudera.
He wasn’t always so ambitious. An Egyptian native, he came to the United States to earn his PhD at Stanford. He wanted to return to Egypt in teach.
“I didn’t have this kind of entrepreneurial spirit. Stanford infects you with it.” In 1999, with encouragement from his Stanford peeps, he put his studies on pause and launched VivaSmart, an e-commerce content management system. He was 30 years old when Yahoo bought VivaSmart for $9 million.
After landing at Yahoo with the buyout, he spent the next eight years there working with Hadoop. Then the startup itch reappeared. In the summer of 2008 and with the help of Accel Partners “Entrepreneur in Residence” incubation program, Accel partner Ping Li introduced Awadallah to Jeff Hammerbacher, who then led the data team at Facebook.
It was not love at first sight.
“Jeff and I knew each other. We worked in a similar field. In some ways I did not like Jeff. One of my key guys at Yahoo left my team and joined Jeff’s team and I was upset that Jeff did that to me,” Awadallah laughs.
He got over it and the two worked on their plans.
Just before they made their pitch Li showed up with another introduction: Mike Olson. Olson was working with Christophe Bisciglia.
Olson was at Oracle, which bought the database company he founded (Sleepycat) in 2006. Bisciglia was running Google’s Academic Cloud Computing program. (Bisciglia has since stepped out of Cloudera.)
Both were hatching similar plans, and Li preferred funding one bigger company than two competitors.
Cloudera started in 2009 with $5 million from Accel. Soon, other big-name investors piled in, like Jeff Weiner (CEO of LinkedIn), Diane Greene (founder and former CEO of VMWare).
With its pedigree, Cloudera began winning business right away.
There is a catch: as an open-source project managed by Apache, Hadoop is free. Cloudera makes its money with Cloudera Manager, a subscription-based Hadoop support service that helps customers manage and troubleshoot Hadoop applications.
Cloudera faces competition, especially from Awadallah’s old pals at Yahoo. Yahoo launched its own Hadoop company, Hortonworks. But Awadallah isn’t worried.
“I told Yahoo when I left, invest in me, this is a huge big thing, but they didn’t see that then. So my thought is now you are three years too late,” he explains.
“History shows that the first mover gets 80 per cent of the market, Red Hat, VMware. Anybody else will get a fraction,” he adds. “We now have the edge in terms of customers across verticals. We’ve got all the early adopters. They will do well as with any new market being created. But they won’t do anywhere as well as us.”