Hadoop, an open source starting point, not the end
Interesting analysis here. Hadoop is a rare case of where open source MapReduce and the Hadoop Fiile System have pioneered the space, with many follow up modifications and systems improving them afterward. The follow up systems have been a result of corporate research or company-initiated projects, such as Cloudera's Impala. In many areas, the Apache project became the gold standard for software in that space (Apache Web Server, Apache Tomcat). In Hadoop's case, it became the springboard that launched a thousand improvements and alternatives.