Photo: Lightspeed Ventures
When it comes to big data, “size doesn’t matter,” Ravi Mhatre, managing director of venture firm Lightspeed Venture Partners just told Business Insider.”It’s not just big data. It’s got to be fast data and it’s got to be meaningful data” he says.
There’s a new wave of startups trying to make it easier to use big data systems. If they succeed, they will really hurt the multibillion-dollar market for business-intelligence software and threaten products like SAP’s Business Objects, IBM Cognos and Oracle Hyperion.
“We’re talking an industry today that’s probably $20 [billion] to $30 billion that I think, overnight, is going to be replaced by a completely new set of platforms,” he predicts.
Big data apps like Hadoop allow companies to store humongous amounts of data on inexpensive servers and storage. As they store more data, it leads to two problems: 1) you can’t easily turn all that data into charts 2) it can take a long time—minutes—to get results when you ask a question, run a query.
For instance, your Hadoop system could store info on every item ever sold by your company. If you wanted to find out what time of day, each month, you sold the most items, you couldn’t output all that into a spreadsheet and start running charts. It could be 10 million rows of stuff.
Mhatre says he’s funded a startup, still in stealth mode, working on tech that will take those 10 million rows and turn them into pictures. “Not so much charts, but other ways a human can visually look at large amounts of data,” he explains.
He didn’t name his stealth startup but another example of a big-data startup is Platfora, funded by Andreessen Horowitz, Sutter Hill Ventures, and the CIA’s investment arm, In-Q-Tel.
Mhatre’s funding other startups that will speed up Hadoop and other big-data technologies. This includes a company called Qubole with a product in beta. Qubole was cofounded by two guys who used to run Facebook’s infrastructure, Ashish Thusoo and Joydeep Sen Sarma.