Haskell simulation of base chain

[Saizan]

Implement and validate a high-level simulation of Praos:

• Based on the work in simulation/ to reuse the TCP and P2P models.
• Produce traces that support measures such as latency of tx adoption block diffusion under different loads.
• Compare with real world data.

First step will be to settle on a reasonable level of abstraction, using the following resources:

• Peras simulation
• Ourobors Consensus specs
• High level description of consensus components in "Introduction to the design of the Data Diffusion and Networking for Cardano Shelley"
• Node-to-Node protocols in "The Shelley Networking Protocol" tech report.

[dcoutts]

such as latency of tx adoption

I think we should avoid modelling txs at least initially, but we can do latency of block diffusion.

[Saizan]

Not even for standalone Praos? Fair, that means we do not need a mempool.

[Saizan]

The Peras simulation works by whole-network diffusion of whole-chains, with a fixed diffusion delay for the chain to become available. In such a model then block diffusion is a parameter of the simulation, not something we can investigate.

[ghost]

@Saizan At the Praos level, I think that's fine. For better fidelity, we could actually not use a fixed delay but something like the distribution from DeltaQ analysis

[dcoutts]

We're currently investigating reusing code from the ouroboros-network-mock package to give us a lot of what we want for free.