Network performance simulation

[josef-widder]

In CometBFT, we are currently working on specifications for several important parts. One motivation is to be able to reduce cost for the users of the software in terms of, e.g., bandwidth, storage, etc. In particular mempool and consensus gossip are responsible for considerable network traffic.

In one of the discussions it was brought up (by @sergio-mena, I guess) that in the future it would be great to use the quint specifications also for performance simulations, e.g., in the spirit of ns-3. We could test out new (hopefully improved) protocols and get some first idea about the impact protocol changes have on performance in realistic settings (as this has not been a concern for the current quint simulator, it does not take timing of events or network traffic into account). So I put the idea here for further discussion.