Hurricane's source code is now publicly available under the Apache 2 license.
What is Hurricane?
Hurricane is a high-performance large-scale data analytics system that successfully tames skew in novel ways.
How It Works
The system employs adaptive work partitioning based on runtime load observations. When nodes become overloaded, they can spawn clones of their tasks during execution, with each clone handling a portion of the original dataset.
This approach allows Hurricane to handle skewed data distributions that would cause significant load imbalance in traditional systems.
Get the Code
Visit the GitHub repository to access the source code and documentation.