In 2006, Doug Cutting joined Yahoo in a team led by Eric Baldeschweiler (also known as eric14 or e14). This team had grid computing experts and users. Eric was in charge of figuring out how to build a next generation search grid computing framework for web searches. Here is a quote from a Yahoo employee at that time that described the situation prevailing at that time:
"Fortunately, and I remember the day well, Eric14 assembled the merry bunch of Grid (then called 'Utility Computing') engineers, and started down the path of rethinking the strategy - focussing on figuring out how to make Hadoop functional, featureful, and robust, instead." (Kumar, 2011)
The new team split out of Hadoop from Nutch with the leadership of Doug Cutting and created an open source Hadoop Framework based upon Hadoop Distributed File System as its storage system, and the MapReduce paradigm as the parallel computing model. Yahoo put more than 300 person-years of effort into Hadoop projects between 2006 - 2011. A team of nearly 100 people worked upon Apache Hadoop, and related projects such as Pig, ZooKeeper, Hive, HBase and Oozie.
In 2011, Yahoo was running Hadoop on over 40,000 machines (>300 cores). Hadoop has over a thousand regular users who use Hadoop for search-related research, advertising, detection of spam and personalization apart from many other topics. Hadoop has proven itself at Yahoo in many revenue driving improvement projects.
Nowadays, Hadoop is a top-level project at Apache Foundation. Hadoop is a software library that contains programs that allow processing of very large datasets, also known as big data, on a large cluster of commodity servers using a simple programming model known as MapReduce. At the time of writing this book, Hadoop 2.7.1 is the latest stable version.
It should be evident from the history of Hadoop that it was invented to solve the problem of searching and indexing massive data sets in large Internet companies. The purpose of Hadoop was to store and process the information inside Yahoo. Yahoo decided to make Hadoop open source so that the Hadoop project could benefit from the innovative ideas and involvement of the open source community.