Sqrrl Raises $2 Million For Secure, Scalable Big Data Technology Originally Developed At NSA

Sqrrl, a big data startup that came out of development at the National Security Agency (NSA), has raised $2 million in seed funding from Matrix Partners and Atlas Venture.

The Sqrrl technology is promised to provide a secure, scalable and easily adaptable way for companies to manage big data.

Here’s the problem with big data that Sqrrl hopes to solve: Much of the data that people want to analyze is under such regulations as HIPAA or Sarbanes-Oxley. You can’t put all the data together without running afoul of regulations. The alternative is to stay in compliance by taking chunks of data that then gets analyzed as a secure data set. But that can be a mess in itself as it causes a fracture with multiple analyses.

A better method, which Sqrrl sees as the opportunity, is to take different elements of a data set and make them secure down to the cell level. A health record, for instance, will have such elements as the age, gender or diagnosis of the individual. With Sqrrl, each element can be inserted instead of the entire health record. Any number of combinations of elements can be used for the data analysis. That provides a richer analysis with infinite potential combinations. The data can be secured to particular groups, further making it regulatory compliant.

Sqrrl is built on Apache Accumulo, a database built on top of Hadoop, the distributed file system. CEO and Co-Founder Oren J. Falkowitz helped develop Apache Accumulo at the NSA. Falkowitz says the NSA has a history of opening its technology.

Apache Accumulo competes with HBase, a database also built on Hadoop for big data analysis. Startups such as Drawn to Scale are building their solutions on HBase.

Drawn to Scale Founder Bradford Stephens said in e-mail:

So Sqrrl is certainly interesting. There’s some nice ideas behind Accumulo that I think are worth merit, and many of them should be implemented in HBase properly. However, Accumulo, like HBase, is not a SQL database — which means that although it may be distributed, it’s extraordinarily difficult for the average developer to figure out how to use. That’s the core problem we’ve solved, and we have a ton of IP behind it.

Falkowitz seems to be taking a similar approach in its market approach. Sqrrl will consist of different products that will be designed to be easily implemented inside an organization using existing tools. It also plans to announce partnerships with business intelligence providers.