alkim0 / tagexec

Introduction

This repository contains the code for Basilisk, a column-oriented database system capable of performing tagged execution as outlined in "Optimizing Disjunctive Queries with Tagged Execution" (SIGMOD 2024). A link to the technical report can be found here. To cite, please use:

@article{kimOptimizingDisjunctiveQueries2024
  author       = {Albert Kim and
                  Samuel Madden},
  title        = {Optimizing Disjunctive Queries with Tagged Execution},
  journal      = {Proc. {ACM} Manag. Data},
  volume       = {2},
  number       = {3},
  pages        = {158:1--158:25},
  year         = {2024},
  url          = {https://doi.org/10.1145/3654961},
  doi          = {10.1145/3654961},
}

Running Experiments

drop_caches

Before, running any experimental code, we must first create a supplemental script for dropping Linux's file system cache. Go into the drop_caches directory, run make, update owner and set the sticky bit:

cd drop_caches
make
sudo chown root drop_caches
sudo chmod u+s drop_caches

Other Scripts

Other scripts are in the scripts directory and managed by poetry. Make sure to run install dependencies with:

cd scripts
poetry install

Generating Data

Generating IMDB Data

To generate the data for the join order benchmark, download the csvs from http://homepages.cwi.nl/~boncz/job/imdb.tgz Then go to the script directory and use the imdb2db script to generate data in a compatible format:

cd scripts
poetry run imdb2db <imdb-csvs-directory>

The output directory may be changed with the -o flag.

Generating Synthetic Data

To generate the data for the synthetic experiments, use the gen-synth-data script:

cd scripts
poetry run gen-synth-data

The output directory may be changed with the -o flag.

Running Experiments

Run experiments with the job-exp and synth-exp binaries:

cargo build --release
./target/release/job-exp
./target/release/synth-exp

The database directory may be changed with the --db-path flag.

Running Tests

To run tests, we must first create the test data. Go into the scripts directory and run:

cd scripts
poetry run gen-test-data

Then in the main directory, run:

cargo test