ABE Cubed

General Info and Versions

The artifact was developed and tested on Ubuntu 22.04.5 LTS with cargo 1.85.1 for the Rust implementation.

The measurements in eval are taken on an AMD server with an Ubuntu VM on 8 physical and 16 virtual cores
and 32 GB of RAM.

Artifact structure

Code

This folder contains the Rust implementation of the scheme with arkworks and criterion benchmarking setup.
The Makefile provides an easy point of entry to build and run things.

Run the tests

Warning: running all tests may take a while (~20 minutes on my laptop)

$ make test

Run the benchmarks

Warning: running all benchmarks may take a while (~3 days on our server)
This produces out.txt files which can be converted to CSV with report_to_csv.jl in eval.

$ make bench

We employ 3 benchmarking strategies, namely 1a, 1b and 2. You can run them individuall using
the make targets bench_01a, bench_01b and bench_02 respectively.

Run the example in main.rs

Info: when successfull, the example has no output/prints.

$ make run

Eval

This folder contains our measured benchmark data and utility scripts to parse output, generate plots
and evaluate the schemes.

• strat01a: the data for benchmarking strategy 01a as presented in the paper
• strat01b: the data for benchmarking strategy 0b as presented in the paper

• strat02: the data for benchmarking strategy 02 as presented in the paper

• report_to_csv.jl: a Julia script to convert out.txt into a CSV file for further processing

• plot_csv.jl: a Julia script to re-create all plots under eval
• *.ipynb: Jupyter (Python) notebooks to perform further evaluations

Scripts

Here are additional scripts used during/for the development of the implementation.

• deduplication.py: an experimental script to model the behaviour of tau/iota functions
• plot_dedup_mappings.py: a script to plot the behaviour of tau/iota for various policies
• generate_test_cases.py: a script to auto-generate test cases for the scheme implementation
• input_gen.py: an experiment script to explore strategies to generate the inputs for our benchmarks