Cloudera Releases SQL-on-Hadoop Tool
Cloudera Inc. yesterday announced the release of its interactive, real-time query engine for Apache Hadoop.
Called Impala 1.0, the open source project has been in public beta since October 2012. "With Impala, users can query data stored in HDFS and HBase directly" Cloudera said. "The framework supports all standard file and data formats available, so users can choose the format that best suits their use case, including the latest in analytics-focused columnar formats like Parquet, and can promote data sharing and reuse across all computing workloads--from batch to interactive SQL--all from a single dataset."
The company said this approach reduces latency found in legacy data warehouse environments and doesn't require having to use proprietary formats or move datasets as it exists with Hive into special systems for analysis. It said Impala SQL is a subset of HiveQL, with some limitations. It supports common Hive interfaces such as JDBC and ODBC drivers and Hue Beeswax.
Impala doesn't use MapReduce. Cloudera said the tool is primarily designed for ad-hoc SQL queries for interactive and exploratory analysis on large datasets, while Hive and MapReduce are preferable tools to use for extremely long-running batch tasks such as with Extract, Transform and Load (ETL) systems. The company said it anticipates Impala also will be used when low latency is needed. It said Complex Event Procesing is ususally done with stream-processing systems. Impala, instead, "most closely resembles a relational database," the company said.
As an Apache-licensed open source project, the Impala platform's souce code is hosted on GitHub and is open to community contributions. Impala can be downloaded via the company's own installation and configuration manager or manually.
David Ramel is editor in chief of Visual Studio Magazine and Application Development Trends Magazine.