
READ ME

1a. Change directory to the project folder  run the command  source CompileChanges.sh

1b. Go to config/single_node_cluster/config

    For steps 3,4,5 follow the current stores.xml
     For steps 6,7 rename stores-new.xml to stores.xml

2. Go to the bin folder, type the following command (start the server in background) 

    sh voldemort-server.sh ../config/single_node_cluster/ >& tmp.log &

    To later stop the server, sh voldemort-stop.sh

Note that the server has to be stopped and restarted again after a change in config(as in 1b) for the new settings to get reflected.

Perform the following steps from the project folder (a11)

3. Run java voldemort/TranslateIntoColumnStore - a simple straightforward implementation of a column store over key value store

4. Run java voldemort/AsynchronousTranslateIntoColumnStore demonstrates how issuing asynchronous gets improves upon 3

5. Run java voldemort/printSubsets to understand how an intelligent vertical paritioning can improve overall response time against a naive column store approach.(Change the path of write.txt according to ur local directories)

6. Run java voldemort/SimpleIndexDriver - an inverted index implementation and performance for a range count query.

7. Run java voldemort/SimpleBucketIndexDriver to see how a bucket index remarkably does better than simple inverted index for range queries


P.S: The important source files are in src/java/voldemort/ and the config files in config/single_node_cluster/config


