Tread on the footprints of predecessors to learn hadoop--structure, focus

HDFs, as a distributed file system, is the foundation of all these projects. The analysis of HDFs is good for understanding other systems. Since Hadoop's HDFs and MapReduce are the same project, we put them together for analysis.

If you take the whole Hadoop as a class in Java, then HDFs is the static variable of this class, and the other projects are the methods in Hadoop.

HDFs

Implementation of Hdfs,hadoop Distributed File system

File system abstraction, which can be understood as a unified file access interface that supports multiple file system implementations

FS

File system abstraction, which can be understood as a unified file access interface that supports multiple file system implementations

IPC

An implementation of a simple IPC that relies on the codec functionality provided by IO

Reference: http://zhangyu8374.javaeye.com/blog/86306

io

The presentation layer. Encode/decode various data for easy transmission over the network

MapReduce

Map/reduce implementation of Hadoop

Filecache

Provides a local cache of HDFs files for faster map/reduce data access

IPC

An implementation of a simple IPC that relies on the codec functionality provided by IO

Reference: http://zhangyu8374.javaeye.com/blog/86306

io

The presentation layer. Encode/decode various data for easy transmission over the network

Tread on the footprints of predecessors to learn hadoop--structure, focus