TALK TRACK Good morning. I’m at Hortonworks. I’m excited to be speaking with you today about how Hortonworks is powering the future of data.
TALK TRACK There are two paths to Hadoop, and you should understand their differences. As the Hadoop platform evolved in the early days, different companies held different philosophies about how to move the open-source technology forward. Some Hadoop vendors took proprietary approaches to Hadoop, filling gaps with extensions that locked customers into their proprietary approaches. This path also limited the pace of innovation to what developers at those companies could build, according to their company’s priorities. The result is typically a disjointed collection of data repositories and apps that don’t work in coordination and that evolve at the pace determined by the vendor. From the very beginning, Hortonworks decided to do Hadoop differently—entirely in the open community, in collaboration with hundreds of the world’s best developers, wherever they may live and wherever they happen to work.
Hortonworks leads the emerging category known as Open Enterprise Hadoop. Hortonworks Data Platform is the Hadoop distribution created and supported by Hortonworks.
Hortonworks Data Platform: Allows you to deploy your cluster in the way that works best for you: on-premises, in the cloud or in a hybrid architecture; on Linux or Windows. It ingests and stores data in its raw form, regardless of its source or format. This schema-on-read architecture is far more flexible than your familiar schema-on-write databases. Without first doing complicated ETL transformations, you can ingest existing tabular data from your RDBMS or newer types of less structured data, such as web clickstream, machine and sensors, social media, mobile, geo-location, or server log data. You can also import your existing structured, tabular data into HDP and enrich it with those newer forms of data. The architecture of Hortonworks Data Platform holds YARN at its center. YARN supports multiple, heterogeneous data access engines that analyze one shared big data set with batch processing, interactive query, search, analysis of streaming data or iterative machine learning approaches. All of these can run simultaneously, centrally managed by YARN. YARN also coordinates cluster resources for enterprise-ready services for cluster operations, data governance and security.
We offer a range of support options for HDP, but that broadly fall into 2 categories: Developer support for the building of applications – for example, assisting customers to develop HDP based applications Enterprise support – again, in a range of options depending on your requirement.