This project is a re-implementation of the map-reduce distributed algorithm in Java.

It is done as a part of the course SLR207-Infrasctructures and platforms for distributed computing at Télécom ParisTech.

the folder scripts contains several scripts to prepare and execute the mapreduce algorithm deploy.sh : deploy slave.jar to the hosts specified in slaves.conf clean_slaves.sh : clean the slaves folders (remove the splits, maps, and the slave.jar files) clean_master.sh : clean the master run_master.sh: run the master program takes two arguments: the file on which to count the words the size of the splits as specified in the split man page

the folder hadoop (in particular the folders hadoop/MASTER/src and hadoop/SLAVE/src) contains the source code for the project

the files slaves.conf contains the list of slaves to use (in the form user@host, one per line)

the folder rapport contains the markdown and pdf versions of the report demanded for the project