Hadoop 2.0 is here, and with it come some big changes. The most notable, as detailed by a recent article from InfoWorld, is MapReduce 2.0, which is now incorporated into a larger system called YARN (Yet Another Resource Negotiator, explained in more detail here). Other changes include the following:
- The JobTracker approach is no more.
- Cross-compatibility with other Apache projects has been expanded.
- Alternatives to MapReduce are available using YARN.
And despite all the changes, backwards compatibility has been retained. In general, Hadoop 2.0 looks to be a lot more flexible than the previous version, which may give it an even more central position in the evolution of Big Data. For a complete overview of the changes, take a look at InfoWorld's full article and see what may be in store for Big Data.