ORCA: Query Optimization as a Service
- Track: HPC, Big Data and Data Science devroom
- Room: AW1.126
- Day: Sunday
- Start: 15:00
- End: 15:25
We all know there is more data than ever before. We do our best to optimize the computation of data, but our tools and techniques haven't kept up. The need for a new approach to query optimization has never been greater.
This motivates the development of ORCA, now a fully open-source query optimizer that is designed to work with any database.
ORCA has achieved a 1000x performance improvement across TPC-DS queries by focusing on three distinct areas: Dynamic Partition Elimination, SubQuery Unnesting, and Common Table Expression.
ORCA is the default query optimizer in the open-source databases, Greenplum Database -- RDBMS data warehouse solution, and Apache HAWQ -- a SQL on Hadoop solution.
Addison will give an overview of ORCA’s architecture, where the project is headed, and how to get involved. The need to rethink query optimization led to the development of ORCA, now a fully open-source query optimizer that is designed to work with any database.
ORCA has achieved a 1000x performance improvement across TPC-DS queries by focusing on three distinct areas: Dynamic Partition Elimination, SubQuery Unnesting, and Common Table Expression. ORCA is the default query optimizer in the open-source databases, Greenplum Database -- RDBMS data warehouse solution, and Apache HAWQ -- a SQL on Hadoop solution.
Addison will give an overview of ORCA’s architecture, where the project is headed, and how to get involved.
We all know there is more data than ever before. We do our best to optimize the computation of data, but our tools and techniques haven't kept up. The need for a new approach to query optimization has never been greater.
This motivates the development of ORCA, now a fully open-source query optimizer that is designed to work with any database.
ORCA has achieved a 1000x performance improvement across TPC-DS queries by focusing on three distinct areas: Dynamic Partition Elimination, SubQuery Unnesting, and Common Table Expression.
ORCA is the default query optimizer in the open-source databases, Greenplum Database -- RDBMS data warehouse solution, and Apache HAWQ -- a SQL on Hadoop solution.
Addison will give an overview of ORCA’s architecture, where the project is headed, and how to get involved.
About the Speaker
Addison is a core member of Pivotal’s Query Processing team. Before building databases at Pivotal, he worked on Big Data applications at Visa Inc. and was a member of Pivotal Labs, Pivotal's agile software development consulting division. He studied computer science at the Georgia Institute of Technology, where he focused on database technologies and machine learning.
Speakers
 |
Addison Huddy |