InfoQ: Columnar Databases and Vectorization

May 27, 2018

This is an in-depth technical article in InfoQ about vectorization in columnar databases and how it relates to Apache Arrow and Dremio.

“Dremio is an open source data processing framework based on Apache Arrow with vectorization capabilities. In this article we discussed some of the key aspects of these capabilities and a detailed tech talk was given at Strata Conference this year in San Jose. In Dremio we make extensive use of Arrow throughout our in-memory execution engine. We recently revamped most parts of Java implementation in Arrow for improved performance and heap usage. Some Tpch queries exhibited as much as a 60% reduction in latency. We have additional enhancements we plan to make, including native SIMD acceleration libraries which we believe will provide significant additional improvements in processing efficiency.”

Get Started Free

No time limit - totally free - just the way you like it.

Sign Up Now

Watch Demo

Not ready to get started today? See the platform in action.

Check Out Demo