Distributed Has Table

The project implements a simplified version of the distributed hash table with MPI in C. A distributed hash table is a decentralized distributed system that provides a logical key-value store where data is distributed across multiple nodes in the network.

The image is taken from hazelcast.com

Installation

Include dht.h in the user program. A sample user program, command.c, is given the source directory.

#include "dht.h"

Compile and run your program. This is an example Makefile:

dht:	dht.o dht-helper.o command.o
	mpicc -o dht dht.o dht-helper.o command.o

dht-helper.o:	dht-helper.c dht-helper.h
	mpicc -c -std=c99 dht-helper.c

dht.o:	dht.c dht-helper.h
	mpicc -c  -std=c99 dht.c

command.o:	command.c
	mpicc -c  -std=c99 command.c

clean:
	rm -f *.o dht 

Commands

There are 5 available operations:

PUT: 0 : stores a value on one of the storage nodes. They key is an integer between 1 and 1000.

GET: 1 : returns the value associated with the key. If not found, -1000 returned.

ADD: 2 : adds the node of the given rank to the cluster, assigning the given id. The data will be redistributed to make up for the addition.

REMOVE: 3 : removed the node with the specified id from the dht. The data is trasnferred to neighboring nodes.

END: 4: terminates the program.

For the examples how the commands are called, have a look at command.c.