Cloudera has announced the general availability of Cloudera Search and an accompanying add-on RTS (Real-time Search) subscription. This product represents the first fully integrated search engine for interactive exploration of data stored in the Hadoop Distributed File System (HDFS) and Apache HBase.
NOTE: Apache HBase is the Hadoop database and as such is a distributed big data store.
Cloudera suggests that "for years now" we have seen databases attempting to provide "search as a feature" in their platforms — although the company says that this approach was largely abandoned in favor of acquiring independent search products that require their own infrastructure, integration, and expertise.
But, and it's a big but, Hadoop's flexibility makes it well suited for search, and consequently, a better general-purpose platform for data exploration than relational databases.
VP of produts at Cloudera Charles Zedlewski explains that Cloudera Search enables natural language keyword searches and "faceted navigation" without additional training or advanced programming knowledge.
Key features include scalable Index Storage in HDFS, batch Indexing via MapReduce for scalable index creation of data stored in HDFS, and HBase comparable to MapReduce. There is also real-time indexing at collection, which makes events searchable as if they are stored in HDFS and HBase through near real-time indexing features, powered by Apache Flume and the Lily HBase Indexer.
"We've taken what was once a relatively complicated and involved freestanding system, requiring its own hardware and operational model, and turned it into a feature of a larger, more ubiquitous open source platform — CDH. We believe this integrated approach represents a big step forward for users of both Solr and Hadoop," said Zedlewski.
NOTE: The RTS (Real-time Search) subscription is a means of getting technical support, legal indemnification, and "continual influence" over the development of the open source projects involved here.
Also featured is simplified field extraction and cross-platform data processing to enable field extraction of any data that is stored in HDFS using optimized Hadoop file formats, such as Apache Avro . Users can "avoid the pain" that many standalone search solutions impose, by promoting reusable configurations and processing activities with the new processing framework, Cloudera Morphlines.
For further reading on Cloudera, see our Hadoop Tutorial Series.