A San Mateo, California, startup called Argyle Data released its flagship anti-fraud software on Thursday, claiming it’s built from the same technologies that let organizations such as Facebook and the National Security Agency analyze their many petabytes of data.
The product, called ArgyleDB, uses the open source Accumulo database technology developed by the NSA to perform deep-packet inspection and create massive databases from that data. Argyle uses Facebook’s open source Presto technology to let users analyze that data using SQL queries and automate future queries against live data. Both technologies are designed to store and analyze data that’s stored in Hadoop, the popular open-source big data platform.
Like nearly every security or fraud-detection company that has launched in recent history, Argyle also uses machine learning algorithms to detect fraud patterns across datasets much too large for humans to make sense of themselves.
But the use of Accumulo and Presto is the really interesting stuff here. Historically, we’ve seen big data technologies such as Hadoop and NoSQL databases emerge from large web companies, get adopted by many other web companies, and eventually then make their way into the mainstream. To see the higher-level technologies that emerged from those earlier projects already acting as the foundations of broadly applicable commercial products seems like an encouraging sign for what’s to come.
A big, scalable open source platform like Hadoop for storing data is great, but big, scalable, open source analytic technologies on top of Hadoop start to open up a lot more possibilities.