BigData needs BigProcessing 🙂 … Hadoop solves one problem, the approach of distributed processing. One of the key benefits is that, the information can be broken down into smaller chunks and processed independently. However, the information must be first stored into HDFS for processing, more suitable for batch-processing, more over data-movement is critical for processing by Hadoop.
In a real-time world, where one needs to processing incoming events (stream) instantaneously, then one will require a different framework. Yahoo S4 is an example of such an alternative framework for event processing.
Refer to the evolving documentation on S4 : http://docs.s4.io/manual/overview.html