Intel is continuing to build out its array of software tools for the Hadoop open-source big data processing framework, with an emphasis on the security and reliability features demanded by large enterprises.
A Data Platform tools suite will become available in the next quarter as a free-of-charge but self-supported Enterprise Edition, as well as a subscription Premium Edition that provides features such as proactive security fixes, regular enhancements and live support.
Intel is competing with the likes of Hortonworks and Cloudera and others in the commercial Hadoop market. The rise of such vendors underscores the fact that Hadoop is “really at a crossroads,” said Jason Fedder, general manager of channels, marketing and business operations for Intel’s data center software division. Linux only took off once companies began investing in hardening its features and also “coalesced around keeping it open,” he added. “The same thing’s happening in the big data domain.”
Customers who choose Intel’s Hadoop distribution over others can benefit from what Intel describes as significant performance improvements thanks to optimizations for Intel’s Xeon processors, solid state storage and networking.
Intel also offers speedier data encryption and decryption within Hadoop with its AES-NI tecnology, according to the vendor. Along with security and reliability upgrades, the Data Platform features capabilities for streaming data processing, iterative analytics and graph processing, according to Intel.
Fedder declined to share how many customers Intel has for its Hadoop distribution. A lot of the work for it began as a lab project in China. To date, most customers are in China although there are users in Europe and the U.S. as well, he said.
Those looking to pinpoint Intel’s intentions for Hadoop should know one thing, according to Fedder. “Where we differ from other players is we’re not trying to build an end-to-end solution,” he said. Instead, Intel wants to be the “operating system” for big data, letting third-party vendors and customers themselves create the application layer on top, he added.
Another problem in the way of mass Hadoop adoption is the availability of programmers and data scientists who can work with the low-level framework. To this end, in the second quarter Intel will also ship a new analytics toolkit aimed at reducing “the complexity, effort, and cost associated with knowledge discovery and predictive modeling,” according to a statement.
The tools will provide generalized algorithms, such as for generating a social graph, that can be applied to various industry use cases, such as retail and financial services.