dhanus / lsm-tree

This is an implementation of a log structure merge tree.

lsm-tree

This is an implementation of a two-level single-threaded log structure merge (LSM) tree. The in-memory data structure is a simple array. The on-disk data structure is also a simple array, which can be sorted using merge sort.

File structure

lsm.c holds the implementation of the LSM tree. test.c holds the test functions. run.py and analyze.py are auxiliary files used to collect and process the data.
create_plots.ipynb is an ipython notebook which takes in the arrays of data as .p files and plots it using NumPy and Matplotlib.

To replicate data collection

Select the test that you want to run from the main function in test.c

make clean 
make 

Select the corresponding test from the run.py file

python run.py 

Select the corresponding test from analyze.py file

python analyze.py 

To replicate plotting

Run the jupyter notebook

jupyter notebook create_plots.ipynb

Re-run each of the cells.