Gossip

The gossip application is an example showing how the Plumtree gossip implementation functions.

gossip is distributed as part of the binary distribution of Tuweni, which you can download from this page

./gossip --help
Usage: <main class> [-h] [--sending] [--numberOfMessages=<numberOfMessages>]
                    [--payloadSize=<payloadSize>]
                    [--sendInterval=<sendInterval>] [-c=<configPath>]
                    [-l=<port>] [-m=<messageLog>] [-n=<networkInterface>]
                    [-r=<rpcPort>] [-p[=<peers>...]]...
      --numberOfMessages=<numberOfMessages>
                            Number of messages to publish (load testing)
      --payloadSize=<payloadSize>
                            Size of the random payload to send to other peers (load
                              testing)
      --sending             Whether this peer sends random messages to all other
                              peers (load testing)
      --sendInterval=<sendInterval>
                            Interval to wait in between sending messages in
                              milliseconds (load testing)
  -c, --config=<configPath> Configuration file.
  -h, --help                Prints usage prompt
  -l, --listen=<port>       Port to listen on
  -m, --messageLog=<messageLog>
                            Log file where messages are stored
  -n, --networkInterface=<networkInterface>
                            Network interface to bind to
  -p, --peer[=<peers>...]   Static peers list
  -r, --rpc=<rpcPort>       RPC port to listen on

You can set up a gossip app to listen with this command:

./gossip -l 9000 -m /tmp/log

In a separate shell, you can send messages to the listener with:

./gossip -l 9001 -p 127.0.0.1:9000 --sending --payloadSize=32 --numberOfMessages=10 -p tcp://127.0.0.1:9000 --sendInterval=1000

This will send 10 messages, of 32 random bytes each. Open the file /tmp/log to see them.

You can create more complex scenarios with multiple gossip listeners, showing the path by which gossip circulates across peers.