There’s a lot of buzz about this one. Silicon Valley smells a megabuck start-up fortune in the making. Pulses are quickening and text messages are flying.
But here’s the zillion-dollar question: Is Clouderareally the next hot open source start-up, perhaps a nascent XenSource (bought by Citrix for a gorgeous $500 million)? Or will it be just another fledgling rushing into the crowded cloud computing sector?
Four really bright tech guys came up with a bright idea. How about if we take Hadoop– a successful open source project – gussy it up, put a corporate face on it, and sell it as a service to companies that need their data mined?
Thus was born Cloudera – note the nod to ‘cloud computing,’ tech’s most fashionable buzzword. (Sometime soon the Starbucks in Mountain View will start selling a Cloud Cappuccino.)
In the software world, Hadoop is the belle of the ball. This open source data processing engine has attracted a bevy of admirers ranging from Yahoo to Facebook to Google to Microsoft.
This relatively new arrival on the scene is brilliantly smart. Hadoop can run petabytes of data across a mass cluster of servers, allowing enquiring minds to sift through mountains of information and find what’s important. In a world awash in data, expect Hadoop to grow ever more popular.
But there’s no guarantee that Hadoop’s data-sifting prowess will translate into commercial success for Cloudera.
Since Hadoop is open source, there’s nothing stopping another competitor from grabbing the code and offering a similar product. For that matter, someone could fork the code and go off in a closely related direction.
But Cloudera has heavyweight backing. It has raised money from venture capital group Accel Partners (though admittedly only a tepid $5 million). And its array of additional investors reads like a list of tech glitterati, including Marten Mickos, former MySQL CEO, and Diane Green, former VMware CEO.
Another challenge for Cloudera: the Hadoop do-it-yourself factor.
Cloudera is using the classic open source business model of selling support and customization, instead of selling shrink-wrapped copies. This model works beautifully for Red Hat, as seen in the Linux provider’s recent robust earnings report.
But Hadoop is a single framework, instead of an entire operating ecosystem like Red Hat provides. Anyone can download Hadoop and install it. While the average person would quickly be lost, in-house IT professionals could (in theory) peruse the Apache guide and plow through.
Does a company really need to hire Cloudera for help?
Here’s where the expertise at Cloudera will need to prove itself. And clearly the group is an A Team. CEO Mike Olson was a VP at Oracle and formerly ran open source database outfit Sleepycat Software; Christophe Bisciglia headed Google’s Academic Cloud Computing Initiative; Amr Awadallah was a VP at Yahoo (he worked with Hadoop at Yahoo) and has a business intelligence background; and Jeff Hammerbacher is a Harvard math wiz and Facebook alum who helped produce Hive, a data warehouse infrastructure built on Hadoop.
If any outfit could convince a big company that it needs to pay for Hadoop consulting-customizing, the Cloudera team is probably it.
Here’s an interview conducted over Skype with Mike Olson, Cloudera CEO: