This document discusses Hadoop 2.0 and its key components HDFS and MapReduce. It provides an overview of HDFS, including its goals of handling hardware failures, streaming data access for batch processing of large datasets, and moving computation instead of data. HDFS architecture is also summarized, describing how it is a very large distributed file system that works with commodity hardware using file replication and recovery from failures. The document then summarizes MapReduce and how it allows parallel processing of data by splitting files into tasks processed across multiple nodes to read data locally instead of over the network.