Recently I picked up the FLINK-36808 bug in the Apache Flink project. The solution to bug covers topics such as SQL LookupJoin, Flink SQL Planner, and Volcano optimizer (which is provided by the Apache Calcite project).

I wanted to explore and deep dive into the Just-In-Time (JIT) compilation and code generation topics. To get some hands-on experience, I was looking for a fun project and came across the Prospero Challenge from Matt Keeter, which was a perfect fit.

Gunnar Morling launched One Billion Row Challenge (1BRC) in the beginning of the year. The goal is to calculate temperature aggregates (min, max, sum) of weather stations. The data is one billion rows of measurements in <string: station>;<double: temperature> format.

These are notes taken to better understand the Java Memory Model, which I now publish as a blog post.

06 August 2018 - Lagos, Portugal