Understanding Cloudera Search

Cloudera Search fits into the broader set of solutions available for analyzing information in large data sets. With especially large data sets, it is impossible to store all information reliably on a single machine and then query that data. CDH provides both the means and the tools to store the data and run queries. You can explore data through:
  • MapReduce jobs
  • Cloudera Impala queries
  • Cloudera Search queries

CDH provides storage of and access to large data sets using MapReduce jobs, but creating these jobs requires technical knowledge, and each job can take minutes or more to run. The longer run times associated with MapReduce jobs can interrupt the process of exploring data.

To provide more immediate queries and responses and to eliminate the need to write MapReduce applications, Cloudera offers Impala. Impala returns results in seconds instead of minutes.

Although Impala is a fast, powerful application, it uses SQL-based querying syntax. Using Impala can be challenging for users who are not familiar with SQL. If you do not know SQL, you can use Cloudera Search. In addition, Impala, Hive, and Pig all require a structure that is applied at query time, whereas Search supports free-text search on any data or fields you have indexed.

How Search Uses Existing Infrastructure

Any data already in a CDH deployment can be indexed and made available for query by Cloudera Search. For data that is not stored in CDH, Cloudera Search provides tools for loading data into the existing infrastructure, and for indexing data as it is moved to HDFS or written to HBase.

By leveraging existing infrastructure, Cloudera Search eliminates the need to create new, redundant structures. In addition, Cloudera Search uses services provided by CDH and Cloudera Manager in a way that does not interfere with other tasks running in the same environment. This way, you can reuse existing infrastructure without the cost and problems associated with running multiple services in the same set of systems.