Open sbellem opened 6 years ago
From @amiller on October 25, 2017 3:5
The experiments might be outdated in design. They're useful for measuring bandwidth usage and other resource costs. To get the high scores we reported in the paper, we used a manually-optimized/compressed encoding format, but tossed it out during refactoring for simplicity. The most realistic experiment is the distributed EC2 experiment. It would be better to run some kind of "docker compose" based local network experiment.
It would be better to run some kind of "docker compose" based local network experiment.
Ah yeah, cool! I could help with that!
UPDATE: Instead of fixing and/or updating the experiments as the original issue intended to do, this issue is now concerned with having a docker-compose based local network.
One of the key goals of this local network is that it will help towards the deployment of a test network.
The local network may also be useful for different kinds of test cases, such as #17.
From @sbellem on October 17, 2017 12:28
The experiments need to be updated to match the changes made in the
dev
branch.NOTE: There's a work-in-progress branch addressing this issue: https://github.com/sbellem/HoneyBadgerBFT/tree/experiments
benchmark tests
README.md
orBENCHMARKS.md
), with steps to reproduce.Copied from original issue: amiller/HoneyBadgerBFT#45