Hortonworks, a contributor to Apache Hadoop, has submitted two new incubation projects to the Apache Software Foundation and also announced the launch of the new “Stinger Initiative.” These three projects seek to address key enterprise requirements regarding Hadoop application security and performance.
Tez Next-Generation Runtime
The Tez incubator project was submitted to the Apache Software foundation to provide a data processing framework as an alternative to MapReduce, to accelerate the performance of data coming in and out of Hadoop, David McJannet, vice president of marketing for Hortonworks, tells 5 Minute Briefing. The Tez proposal provides an advanced runtime built on Hadoop YARN that is expected to improve the latency and throughput of Hadoop applications.
Hadoop Gateway for Security
The second incubator project for the Hadoop Gateway allows for a simplified security mechanism for accessing data across the Hadoop infrastructure, says McJannet. This proposal deals with the need for a single point of authentication and secure access for Apache Hadoop services in a cluster, streamlining security for Hadoop.
Stinger Initiative to Accelerate Apache Hive
The Stinger Initiative is a broad initiative to accelerate Apache Hive, which is the primary SQL-oriented interface for getting data in and out of Hadoop to make it much, much faster, explains McJannet. “We believe that we can make Hive up to 100 times faster than it is today. The Stinger Initiative represents a concerted effort by Hortonworks and a whole host of other contributors from other organizations to make Hive faster.”
Because roughly 50% of Hadoop users depend on Hive for SQL-based operational data processing, achieving enhancement of Hive’s SQL capabilities and optimizing its query performance in support of user-focused SQL interactions is a critical need, says Hortonworks. This initiative will assure that Hive remains the de-facto standard for SQL queries with Hadoop.
Hortonworks' Approach to Accelerate Hadoop Adoption
The approach Hortonworks is following with these projects is very consistent to the approach the company has followed for some time – to identify requirements, articulate them to the community, and then either lead the projects there or incubate new projects to further the technology, says McJannet,
“We believe we can accelerate adoption of Hadoop in general by staying very consistent with this 100% open source approach rather than taking the work done in the open source community and selling a proprietary distribution of it and monetizing it with management tools or proprietary interfaces,” he notes. “The ecosystem and the market will mature fastest if everybody is operating off a common distribution which is 100% open source with an avowed intent of enabling the ecosystem rather than competing with it.”
For more information, visit http://www.hortonworks.com.