- Install the Nix Package Manager.
- Enable Flakes.
- List available packages with:
nix flake show github:CRYPTO-KU/FaultTolerantSmartGridAggregation
- Get a package in a temporary shell:
nix shell github:CRYPTO-KU/FaultTolerantSmartGridAggregation#aggft-simulate # Now you can run `aggft-sim`
You can use and extend the core AggFT implementation that is provided as a Python package. Or you can use the companion simulation package to run simulations.
There are three steps to run a simulation with aggft-sim:
- Write simulation spec.
- Run simulation.
- Analyze results.
Documentation for the spec format is available at docs/spec.md. And documentation for the output format is available at docs/output.md.
There is a special simulation command to run simulations based on the topology
of Figure 1 in the paper. The command is aggft-sim-fig. Documentation for
its spec format is available at docs/fig-spec.md. It has the
same output format as aggft-sim.
The simulations are not deterministic. There are three reasons for that.
The first reason is that cryptographic keys are not seeded. This means that simulations using the same spec will have different cryptographic keys. While this should not effect the simulation results, it should be noted.
The second reason is the use of multi-threading and multi-processing. Thread or process scheduling can have an effect on the results.
The third reason is the use of time. This AggFT implementation relies heavily on time. Running the simulations on a slow machine could cause more timeouts and change the results.
However, the first and second reason should not greatly impact the results. To tackle the third reason, the simulation package reports any issues with timouts in the generated results. If you have such issues, you can run smaller-scale simulations, or relax the time limits.
Results in the results directory are CSV files compressed with 7z. You can uncompress them using:
7z x $FILENAME