Can an open-source framework for the storage and large-scale processing of data on clusters of commodity hardware help an organization save the millions of dollars it spends on traditional databases? Venture capitalists believe so, as they are pouring money into tech startups that will enable that change.
The open-source framework is Hadoop, a Big Data management tool, that, as we wrote in 2011, “breaks down huge data sets into small segments that can be run by applications more quickly and cheaply in virtualized (cloud) database environments.” The data is then run through analytic tools. The developers of those tools are the ones attracting the money. As the Wall Street Journal reported on Thursday, Hortonworks closed a $100 million funding round this week, and VCs infused both Cloudera and Platfora with more capital last week.
In addition, Intel is reportedly ditching its own Hadoop distribution and is going to support Cloudera’s version of Hadoop, according to VentureBeat. Intel Capital may also invest in Cloudera.
“The flood of cash from VCs to these startups reflects the view that customers are ready to move on from the current generation of computing, typified by expensive licensing and restrictive computing options to a new, more flexible approach based on commodity hardware and a more open-ended approach to computing,” the WSJ wrote, quoting Paul Maritz, CEO of Pivotal, a software firm.
The former chief technology officer of Sears Holdings said companies are finally implementing Hadoop, cutting software licenses and hardware maintenance fees, especially since software developers, including Facebook, are releasing open-source tools that speed up queries on Hadoop, according to the WSJ. A company started up by the former chief technoloy officer of Yahoo, Altiscale, is developing a version of Hadoop that runs as a service in the cloud.
Online trading firm tradeMONSTER Group is one company that is looking to pull the plug on its existing online analytics processing system. The company’s CTO told the WSJ that its existing system isn’t robust enough to process years of customer behavior data.
Image: Intel Free Press, Wikimedia Commons, CC BY-SA 2.0