Apache Hadoop
Doug Cutting released an open implementation of the file system and MapReduce, and what only Google could do became available to anyone.
Why it matters
The gap between those who could process web scale and everyone else shrank to the price of a cluster of old machines.
The project is named after the author's son's toy elephant. Hadoop started a whole industry that came to be called big data, and a generation of engineers who could work with a cluster. Its limitation was speed: every step wrote intermediate results to disk. Spark solved exactly that in 2012.