The engine is an enhancement to Cloudera's Big Data platform, known as Cloudera Enterprise. In describing the query-engine's uniqueness, Cloudera claims this is the first time both real-time and batch operations are available for unstructured and structured data in one massively scalable system.

Cloudera offers a commonly used version of Hadoop, an open-source data framework designed for handling Big Data.

In its announcement, Cloudera said that the new query engine will enable organizations to "process data at petabyte scale and, on the same system, interact with that data in real time to deliver 'speed-of-thought' insights." In short, the company said, the new tool will allow organizations to "ask bigger questions" of their data.

SQL Queries

The Apache-licensed, open-source query engine, Cloudera Impala, is specifically designed for real-time query of data stored in a Hadoop Distributed File System (HDFS) and in HBase, a non-relational distributed database. Interactive queries for Impala can be expressed as SQL.

The company said that Impala operates 10 times as fast as the existing Hive/MapReduce, and can be even faster, depending on the workload. It pointed to cost savings for analyzing Big Data with real-time queries, by using this open-source technology with commodity hardware.

Cloudera said that, in a recent survey it conducted of more than 100 customers, over 70 percent were looking at how to extract value from Big Data. Operational IT efficiency and competitive advantage were cited by the customers as reasons for adopting Hadoop, but the vast majority also indicated they needed faster methods of querying than the batch operations that had been available.

'Most Exciting' Since Hadoop

In its announcement, the company pointed to one of its clients, travel Web site Expedia, which said that it uses the Cloudera Enterprise platform to manage more than 4 petabytes of data. With Impala added, Expedia said the enhanced Enterprise Real-Time Query platform allows the creation of one single platform for Big Data, instead of having to maintain several systems for archiving, extracting, transforming, loading, and analytics.

Cloudera CEO Mike Olson said in a statement that, "until now, enterprises had to limit the work they did with Hadoop because batch-mode processing using MapReduce was just too slow for some business problems." Impala, he explained, will enable organizations to store all their data in Hadoop and "use the same hardware to do both powerful analytics and run real-time queries using industry-standard tools and the SQL language."

In fact, Cloudera co-founder and Chief Scientist Jeff Hammerbacher characterized Impala as "the most exciting open-source project since Hadoop," adding that it was "the most important framework beyond MapReduce for analyzing data stored in HDFS and Hbase."