One of the killer applications that gets people excited about Big Data in general is being able to analyze massive amounts of machine data. After all, all that data is there for the taking; prior to the advent of data management frameworks such as Hadoop there was no real way to cost-effectively store all that data.
Now the folks at Concurrent want to make it easier to actually build applications using machine data. The company this week released Patterns, an open source scoring engine that allows analysts and data scientists to create applications that will run on top of Hadoop.
Concurrent CTO Chris Wensel says Patterns runs on top of Cascade, a framework that Concurrent developed to allow developers to use SQL to create Big Data applications running on Hadoop. Patterns works by allowing an analyst or data scientist to invoke the Predictive Model Markup Language (PMML) or a simple programming interface to create an application capable of querying Hadoop.
One of the reasons that enterprise IT organizations have been slow to embrace Hadoop is the complexity associated with building applications. Analysts and the IT staff both need to learn new skill sets. Wensel says Patterns is designed to allow organizations to extend the skillsets they already have to Hadoop.
The end result, says Wensel, is a lot of freedom to innovate using machine data that previously was inaccessible to most organizations.
As the ecosystem around Hadoop continues to expand and mature, it looks like we’re getting closer to being able to build and construct Big Data applications with the people and skills we have, versus the ones we don’t and might never find.