Distributed state machine

Distributed state machine is a project to compare the performance of multiple consensus algorithms.

The goal of a distributed state machine is to have multiple instances (servers) holding the same log

The client can query any server (also called node) to get the current version of the log. The client can also request to add something in the log.

The system is supposed to be fault-tolerant, which means that if some nodes happen to crash, the client should still be able to communicate with functionning nodes.

Context

This project is part of my end-of-studies research project on consensus algorithms.

I started my work with a bibliographical survey and I also did before this a fault-tolerant decentralized calculator.

Unfortunately, the decentralized calculator is not a suitable context to implement the Snowball consensus protocol, which I wanted to study.

Let's compare some consensus algorithms, in the context of building a distributed state machine

Versions

• master branch implements the Raft consensus algorithm.
• slush, snowflake and snowball branches implement the different iterations of the Snowball consensus algorithm

Testing

Select your parameters in config.json (Parameters are different depending on the branch you are on and the consensus algorithm that you are using) :

{
    "updateSystem" : If true the systems remap when one node is failing completely (failing to reply to request within 200ms)
}

Launch the containers with :

bash launch.sh <3 4 5 6>
# number of containers to start (min 3 max 99)

Add one element to the log by querying one container with :

curl -X POST localhost:800<1 2 3>/add-log -H 'Content-Type: application/json' \
    -d '{"position":10, "content": "Foo"}' # ask container <1 2 3> to append "Foo" as log number 10

Request one element to the log by querying one container with :

curl -X POST localhost:800<1 2 3>/request-log -H 'Content-Type: application/json' \
    -d '{"position": 10}' # ask container <1 2 3> for the content of log number 10

Check the logs of container <1 2 3> with :

docker logs --follow sm-server-<1 2 3>

You can also try crashing containers with :

docker stop sm-server-<1 2 3>

See what happens in the logs as the consensus algorithm run !