Evaluation: Efficient Updates for Worst-Case Optimal Join Triple Stores

Paper Data

All measurements, tables and figures used for the paper are in the data.zip archive.

The technical report for the algorithm is provided as hypertrie_update_technical_report.pdf. It also contains a table that compares offline bulk loading speed and storage efficiency of the tested systems.

The proof of time complexity of Hypertrie updates is provided as hypertrie_update_proofs.pdf.

Due to file size limitations, some files cannot be hosted anonymously. Therefore, all links in this repository have been anonymized to maintain double-blind review standards. Upon acceptance, we will provide the original links.

Requirements

Following executables need to be installed:

• docker
• bash
• curl
• wget
• zstd
• (python 3.13)

Add ulimit -n 64000 to your .bashrc. This increases the number of open files allowed by the system. Log out and in again to apply the changes.

Initializing the Environment

We recommend to install python 3.13 through pyenv. To do so, run the following commands:

curl https://pyenv.run | bash
export PYENV_ROOT="$HOME/.pyenv"
[[ -d $PYENV_ROOT/bin ]] && export PATH="$PYENV_ROOT/bin:$PATH"
eval "$(pyenv init -)"

Then, install python 3.13:

pyenv install 3.13 # make sure this is successful. If not, install the required dependencies from your package manager, run again and confirm [Y] if asked to continue because it already exists. 

Afterward, check out the repository and set up the environment:

git clone <this-repo>
cd <this-repo>
pyenv local 3.13
pip install -r requirements.txt

Run the Experiments

Navigate to the repository root, ensure that the correct venv or pyenv is activated and run the following:

python run_benchmarks.py --tripletstore tentris-baseline tentris-insdel graphdb virtuoso oxigraph --dataset dbpedia wikidata

This will run the following steps:

1. download or build the triple stores to benchmark_workspace/triplestores

2. download the benchmarks to benchmark_workspace/benchmarks

3. download the benchmarking tool to benchmark_workspace/mini-iguana

4. Runs the benchmark for each triple store and dataset:

   1. load the dataset
   2. run the warmup queries
   3. run the update changesets

   Results are stored to benchmark_workspace/results

Generate Tables and Figures

Navigate to the repository root, ensure that the correct venv or pyenv is activated.

Make sure the requirements are installed:

pip install -r requirements_eval.txt

Run the following command to generate the tables and figures:

python generate_tables.py
python generate_figures.py

Tables are stored to tables and figures to figures.