- Big Data Hadoop, Map Reduce
- HDFS (Hadoop Distributed File System) to store data on nodes.
- MapReduce to process data on the nodes.
- Files are split into 64 Mb blocks and each block is stored on a separate node.
- HDFS creates 3 copies of each block for redundancy.
- There is a namenode holding the metadata of where the files are split and duplicated.
- There can also be a stand by copy of the namenode to avoid having problem if the active namenode goes down.
- Volume, Variety, Velocity (= Generating and recording a lot of data, in various formats very fast)