A little over a year ago, we introduced monad-viz, A Visual Representation of MonadBFT. For a fixed network configuration, this helped build intuition about validator state and messages over simulated time by running Monad’s consensus inside your browser, reducing the required dependencies to software that ships with every modern device.
Since the first version of monad-viz, significant improvements have been made to MonadBFT itself, with a next-generation consensus protocol, Cadence, well under development. During a recent hackathon, we explored upgrades to monad-viz, in particular, we wanted to tweak the network’s conditions on the fly and observe the effects without rebuilding the application.
The new debugger turns the previous animation loop into a small network laboratory. While the simulation is running, users can:
- Change the latency between individual validators (drag and drop supported)
- Drop messages in either direction
- Create and remove network partitions
- Take validators offline and bring them back
- Inspect consensus messages
- Track the finalization progress of pending blocks
Below is a miniature, mobile-friendly demo, but we highly encourage use of the full version, which exposes the more powerful network configuration tools. In addition to the above on-the-fly uses, we’ve added support to reset the initial validator configuration (number of validators and associated stake-weight).
Network changes are recorded at the current simulation tick. Moving backward or forward through time deterministically reapplies those changes, making it possible to inspect how a particular failure scenario develops. As Cadence development progresses, tools like monad-viz will evolve with it. More to come!
