With its recent release, Apache Spark has promoted Cornet—the BitTorrent-like broadcast mechanism proposed in Orchestra (SIGCOMM’11)—to become its default broadcast mechanism. It’s great to see our research see the light of the real-world! Many thanks to Reynold and others for making it happen.

MLlib, the machine learning library of Spark, will enjoy the biggest boost from this change because of the broadcast-heavy nature of many machine learning algorithms.