Skip to content

Simulator for some PoS or consensus algorithms. Includes dfinity, casper IMD

License

Notifications You must be signed in to change notification settings

nkeywal/wittgenstein

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

CircleCI

Wittgenstein

A simulator for some PoS, consensus or just distributed algorithms. Includes dfinity, Ethereum's Casper IMD, San Fermin and others. A somehow outdated presentation can be found here.

It allows you to test protocols, and to generate large graph and videos of the protocol in action, such as this one (generated for Handel: aggregation at scale with Byzantine nodes (click on the image to see the video): Watch the video

Why this name?

Wittgenstein was a concert pianist. He commissioned Ravel's Piano Concerto for the Left Hand, but changed it: 'lines taken from the orchestral part and added to the solo, harmonies changed, parts added, bars cut and at the end a newly created series of great swirling arpeggios in the final cadenza.'

As it's often what happens with mock protocol implementations, it looked like the right name.

How to build it

You will need java 9+ and gradle installed.

To check everything is correct:

gradle clean test

You can build a jar with gradle:

gradle clean shadowJar

How to run it

Once built:

java -Xms6000m -Xmx12048m -classpath protocols/build/libs/wittgenstein-all.jar net.consensys.wittgenstein.protocols.GSFSignature

This command is typically for a 16GB machine. The memory is very important when you want to simulate tens of thousands of nodes. If you have less memory, use lower values for -Xms and -Xmx, and run the simulations with less nodes.

But you're actually supposed to write code to implement your specific scenarios today. An obvious improvement would be to be able to define scenarios reusable between protocols.

How to implement a new protocol

For a complete description go to the wiki.

Here is an example:

public class PingPong implements Protocol {
  /**
   * You need a network. Nodes are added to this network. Network latency can be set later.
   */
  private final Network<PingPongNode> network = new Network<>();

  /**
   * Nodes have positions. This position is chosen by the builder.
   */
  private final Node.NodeBuilder nb = new Node.NodeBuilderWithRandomPosition();

  /**
   * Messages, exchanged on the network, are specific to the protocol.
   */
  static class Ping extends Network.Message<PingPongNode> {
    @Override
    public void action(PingPongNode from, PingPongNode to) {
      to.onPing(from);
    }
  }

  static class Pong extends Network.Message<PingPongNode> {
    @Override
    public void action(PingPongNode from, PingPongNode to) {
      to.onPong();
    }
  }

 
  class PingPongNode extends Node {
    int pong;

    PingPongNode() {
      super(network.rd, nb);
    }

    void onPing(PingPongNode from) {
      network.send(new Pong(), this, from);
    }

    void onPong() {
      pong++;
    }
  }

  @Override
  public PingPong copy() {
    return new PingPong();
  }

  @Override
  public void init() {
    for (int i = 0; i < 1000; i++) {
      network.addNode(new PingPongNode());
    }
  }

  @Override
  public Network<PingPongNode> network() {
    return network;
  }


  public static void main(String... args) {
    PingPong p = new PingPong();
    
    p.network.setNetworkLatency(new NetworkLatency.NetworkLatencyByDistance());

    p.init();
    PingPongNode witness = p.network.getNodeById(0);
    p.network.sendAll(new Ping(), witness);
    for (int i = 0; i < 1000; i += 100) {
      System.out.println(i + " ms, pongs received " + witness.pong);
      p.network.runMs(100);
    }
  }
}

This will print:

0 ms, pongs received 0
100 ms, pongs received 38
200 ms, pongs received 184
300 ms, pongs received 420
400 ms, pongs received 765
500 ms, pongs received 969
600 ms, pongs received 998
700 ms, pongs received 1000
800 ms, pongs received 1000
900 ms, pongs received 1000

About

Simulator for some PoS or consensus algorithms. Includes dfinity, casper IMD

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 4

  •  
  •  
  •  
  •