Raftify - Highly Available Validator Clusters

Hello, fellow Cosmonauts!

With the Stargate testnet right around the corner, this is a great time for us to draw attention to our high availability solution Raftify.

In a nutshell, Raftify is an implementation of Raft’s leader election algorithm which enables the creation of a self-managing cluster of simultaneously running validators using the same key pair to sign and broadcast messages into the blockchain network. We have made quite a few modifications to the leader election algorithm to ensure double-signing protection, leader stickiness and a single leader during arbitrary network partitions.

Key Benefits

  • A minimal setup requires only three nodes
  • No log replication, thus reducing communication overhead between nodes
  • Robust double-signing protection, no leader flip-flopping and guaranteed single (or no) leader during network partitions
  • Dynamic failure tolerance (adjusts automatically to the current cluster size)

We have been working on Raftify for quite some time now and achieved satisfactory stability and robustness. Now, given that Raftify is feature-complete with v0.3.0, we’d love for the community to try it out and put it through its paces. We think Raftify might be interesting for those of you who don’t already have a good strategy for high availability in place, or for those of you who are already using a Raft-based active/active setup and would like to try out a more cost-effective alternative with reduced communication overhead.

If you would like to read more about Raftify and how it works, you can check out our Medium article. The code for Raftify is open-source, so you’re welcome to check it out on GitHub.

If you find any issues or bugs, please file an issue in the Raftify repository.

Have fun playing around with Raftify!