Big Data gives us access to unprecedented amounts of information. Unfortunately most organizations aren’t able to use all the data they have at their disposable simply because formatting it for consumption would be too resource-intensive (in terms of man hours and expensive data warehouses).
Hadoop is a low-cost, extensible option for semi-structured and unstructured data, but you still have to get the data out.
Apache Drill is a low-latency query engine that lets end users run SQL-like queries directly against live data as it arrives in Hadoop, even on top of schema-less or nested sources.
And Drill exposes JDBC/ODBC interfaces so you can leverage the BI tools you’re already familiar with. And even mash that data with information from more traditional sources.
MapR put together a great demo using Lumira to combine sensor data with packet captures to understand a failing sensor. In the demo Drill is used to query JSON sensor data in a CSV file directly and mashing in data from wireshark. The source of failure becomes obvious when visualizing the data with SAP Lumira.