Description

This is the implementation for the distributed quantum consensus protocol from Ben-Or and Hassidim [1] for the synchronous model that can reach agreement in a constant expected time even in the presence of a full information, adaptive, and computationally unbounded fail-stop adversary that stops up to $t < \frac{n}{3}$ processes. This protocol is simulated under Qiskit runtime local testing mode with Qiskit Aer (V2) Sampler.

Set up

In order to run this implementation you must install the following libraries in the environment that will run it:

pip install ipynb qiskit qiskit[visualization]

Organization

The agreement protocol is the main program to run and can be found under fail_stop_agreement.ipynb. This protocol uses the weak global coin protocol implemented under weak_global_coin.ipynb. After an execution of the whole agreement protocol the tests of agreement, validity, and termination that must be fulfilled by a consensus algorithm (and a lemma that is used to prove the above properties) will be run. These tests are found in protocol_tests.py. The processes taking part of the protocols can be stopped by an adversary, its implementation can be found in adversary.ipynb. Finally, the only parameter that might be changed is the number of processes n¹, which can be found in the globals.py file, together with other global variables.

The visualization.ipynb file contains the neccesary code to plot the circuits and histograms used in this implementation.

References

[1] Ben-Or, M., & Hassidim, A. (2005, May). Fast quantum Byzantine agreement. In Proceedings of the thirty-seventh annual ACM symposium on Theory of computing (pp. 481-485).

Footnotes

1. Note that n cannot excede 384 and with $n = 244$ processes it already takes on average 40 min to execute the whole agreement protocol. ↩