BlueTalon brings Hadoop security down to the file system

29.09.2015
Big data can mean big threats to security, thanks to the tempting volumes of information that may sit waiting for hackers to peruse. BlueTalon hopes to tackle that problem with what it calls the first-ever filtering and dynamic masking capabilities for use directly on the Hadoop Distributed File System (HDFS).

Though Hadoop has enjoyed widespread popularity for big data purposes, the technology's core didn’t originally contain any security controls, said Nik Rouda, a senior analyst with ESG Research.

"It’s only been recent years when the main distribution vendors -- Cloudera, IBM, Hortonworks, MapR -- starting to add this functionality as a prerequisite for adoption by enterprises," Rouda said.

There's no single common approach or security framework; rather, the market is dominated by a wide variety of partial solutions, he said.

By offering fine-grained access control on HDFS, the storage layer used by all Hadoop components, BlueTalon says its approach gives enterprises a "catch-all security blanket" that not only blocks users from bypassing security controls to access data directly in HDFS, but also eliminates the risks associated with security policies that are siloed in specific applications

BlueTalon is demonstrating its technology this week at the Strata + Hadoop conference in New York. Last month, Cloudera certified the BlueTalon Policy Engine for integration with its Apache Hadoop distribution.

"Blue Talon is interesting in their flexibility and control and applicability to multiple environments," said Rouda.

It's still early days for the company, but as security, privacy and governance issues continue to gain attention, "the market will need to find solutions that fit," he said. "I like their architecture and policy-driven design."

Katherine Noyes