Why We Built InfluxDB 3's Query Engine on Apache DataFusion

Session Date: May 26, 2026
Time: 8:00am (PT) | 3:00pm (GMT) | 4:00pm (BST)

In this webinar, we show how DataFusion powers InfluxDB 3’s query layer. Built on the modern FDAP stack (Flight, DataFusion, Arrow, and Parquet), InfluxDB delivers high-performance SQL queries, full JOIN support, and a continuously improving engine shaped by the broader open source community. We will discuss why the legacy approach of building a vertically integrated database is expensive and error prone, and how this leads to customer pain such as SQL “dialects”, custom file formats, data lock in, and a fragmented ecosystem.

What we’ll cover:

• Live technical demo: DataFusion powering InfluxDB 3 queries, including JOINs, aggregations, and high-cardinality workloads.
• The ecosystem advantage: How investing in shared open -source infrastructure delivers outsized returns for InfluxDB customers.
• What’s new: Recent DataFusion releases and what they mean for your InfluxDB query performance.

Andrew Lamb

After spending many years as C/C++ systems programmer (databases and compilers), and a stint working on Machine Learning startups (as one does), Andrew now works at InfluxData with Paul Dix and a talented team of engineers on InfluxDB 3, a new engine for time series data.