Skip to content

cskiraly/das-simulator-nim

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

39 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Data Availability Sampling (DAS) Simulator - Nim edition

Author: Csaba Kiraly, Codex team

The goal of the simulator is to study and develop techniques for DAS (Data Availabilty Sampling). The simulator is composed of

  • the Shadow Network Simulator: open-source framework to run real application code over a simulated network
  • nim-libp2p: an experimental version of nim-libp2p code for DAS
  • custom DAS application logic

This is our second simulator, the first being written in Python and available in the DAS Research repo.

The two simulators have different levels of abstraction and serve different purposes:

  • the Python simulator focuses on abstraction, scale, and ease of modification;
  • this one, written in Nim, is much closer to a real implementation.

Both simulators aim to serve the development and comparative analysis of different DAS variants, including full DAS (with 2D encoding and blocks in the range of 32MB), but also PeerDAS, SubnetDAS and other variants.

Acknowledgements

The Shadow Network Simulator was originally developed for the Tor project.

nim-libp2p is in development and production use by the Nimbus team.

The code of this library is based on the original work of the Vac team.

This work is supported by grant FY22-0820 on Data Availability Sampling from the Ethereum Foundation.

Techniques and protocols supported

While many things are still work in progress, the simulation already has a large number of features implemented:

  • configurable block size
  • configurable segmentation, with optional 1D or 2D erasure coding (simulated)
  • column/row based diffusion using GossipSub(-like) protocol
  • cross-forwarding between rows and columns
  • erasure coding based reconstruction (simulated), with forwarding and cross-forwarding
  • DAS specific Gossipsub modifications (batch, rarest first, randomized order dispatch)

Features in development

  • ID-based custody and/or interest
  • timout-based sampling protocol
  • LossyDAS, IncrementalDAS and DiDAS support
  • more Gossipsub modifications (duplicate supression, etc.)
  • custom Diffusion protocol
  • custom protocol unifying diffusion and sampling

Usage

Install Shadow

Follow instructions for dependencies and for installing the Shadow.

Install Nim and Nimble

Follow instructions to install Nim.

Nim dependencies

Most dependencies will be installed in the nimbledeps folder, which enables easy tweaking. nim-libp2p is included in the shadow/vendor folder as a git submodule.

nimble install -y [email protected] #workaround for bad nimble version resolution
nimble install -dy

Run the simulator

cd shadow
./run.sh <runs> <FastNodes> <SlowNodes> [PacketLoss(float0..1)]

About

DST-shadow

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Nim 55.8%
  • Awk 38.5%
  • Shell 5.7%