MapR Technologies has introduced the next major release of the MapR Ecosystem Pack program, a broad set of open source ecosystem projects that enable big data applications running on the MapR Converged Data Platform while ensuring inter-project compatibility. These latest enhancements also add flexible access and provide new capabilities for streaming applications.
“We’re always looking to give our customers immediate access to cutting-edge tools they need to be successful in their big data deployments,” said Will Ochandarena, senior director, product management, MapR Technologies. “Spark and Drill continue to be two of the most widely adopted ecosystem projects, and this release makes them even easier to adopt for production use.”
The MapR Ecosystem Pack removes the complexity of coordinating many different community projects and versions. MapR develops, tests, and integrates open source ecosystem projects such as Apache Drill, Spark, Parquet, Hive, and Myriad, among others. The new MapR Ecosystem Pack version 2.0 now includes:
- Support for the Kafka REST API and Kafka Connect, opens up new ways to access event data in MapR Streams. The Kafka REST Proxy for MapR Streams lets customers use any development language in any environment that supports HTTP to work with streaming data. Kafka Connect for MapR Streams delivers a framework for standardised access between MapR Streams and the most popular data sources and targets. These capabilities further enable customers to build IoT-scale, global systems of record with MapR Streams by allowing embedded devices like microcontrollers to produce and consume data in real time using REST, while integrating data with other systems like RDBMSs and search engines.
- Support for Spark 2.0.1 adds new features such as whole stage code generation that make programs run faster and thus deliver quicker results. Also, the in-memory columnar feature stores data in an optimised format in RAM to allow faster analytical queries.
- Low latency queries, optimised BI experience, and dynamic UDFs come to Drill 1.9. Key improvements speed up large scale I/O intensive analytics queries up to 33% and advanced filtering and pushdown capabilities reduce I/O by up to 70% for TPC-H queries. The new release enhances metadata query performance and introduces flexible JOIN syntax that optimizes Drill usage with industry standard BI tools.
- MapR Installer Stanzas enable API-driven installation of MapR clusters on-premises or in the cloud. Part of the Spyglass Initiative, this feature helps users build a Stanza, which is a configuration file that describes a cluster and executes it programmatically to automate new deployments. This is especially useful for quickly deploying elastic clusters across the cloud.
