We’re pleased to announce general availability of Starburst Presto 208e! In this release, our major focus was to bring enterprise security to Presto. We know it was long awaited, and now it’s here!
Here’s what’s new:
Presto Security
Role-based access control (RBAC)
- Presto integration with Apache Ranger and with Apache Sentry for role-based access control (RBAC)
- Just point Presto at your existing access control policies in Apache Ranger or Apache Sentry to secure access to your data warehouse or data lake
- Includes Column-level access control
- Support for Apache Ranger column masking is available as a feature preview
Data encryption at rest and in motion
- Data encryption at rest via support for HDFS encryption zones (Hadoop client-side transparent encryption, a.k.a. Hadoop KMS)
- Data encryption in flight via support for HDFS wire encryption, which protects data moving to and from HDFS.
Oracle Connector for Presto
The Oracle Database connector makes it easy to integrate data from Oracle databases into your Presto SQL platform, and perform fast, interactive SQL queries. With this connector, you can execute queries that draw data from one or more Oracle database instances, or from Oracle and non-Oracle sources (e.g., S3 buckets, HDFS files, NoSQL stores, other SQL databases, etc.).
This is arguably the easiest and most economical way to integrate enterprise data from Oracle into new Presto data warehouses and data lakes.
Avro Schema Support in Presto Kafka & Hive Connectors
Presto’s Hive and Kafka connectors now support the Avro schema evolution feature with backward compatibility. With backward compatibility, a newer schema can be used to read Avro data created with an older schema. Newly added/renamed fields must have a default value in the Avro schema file.
Even Faster Presto
We improved memory usage and other things to help improve performance with:
- Scalar correlated subqueries with streaming aggregations (kudos to Maria Basmanova @ Facebook for her work on this)
- Correlated subqueries that require coercions
- Support for gathering basic and column statistics when writing data in Hive connector (thanks to Andri Rosa for contributing to this!).
Presto Stability Improvements
On top of the new features mentioned above, the new release brings important stability fixes to Presto 0.208, including:
- Fix data loss when writing bucketed sorted tables (thanks to David Philips for his work on this).
- Eliminated a broadcast join bug that caused frequent JVM full garbage collection pauses and rare JVM OOM crashes.
- Greater resilience when Presto reads table statistics generated by Apache Impala.
- Extend predicate pushdown for DECIMAL and CHAR types in JDBC connectors.
If you would like more detail, please read the Presto 208e release notes.