Groot

I am Groot.

Usage

Groot provides a distributed KV store for ephemeral data. It utilizes LWW-register CRDTs and Hybrid-Logical clocks to ensure availability and a level of consistency. For technical information on Groot's implementation, please refer to the docs.

Installation

def deps do
  [
    {:groot, "~> 0.1"}
  ]
end

Usage

# Changes are propogated to other nodes.
:ok = Groot.set(:key, "value")

# Read existing values. "Gets" are always done from a local ETS table.
"value" = Groot.get(:key)

set operations will be replicated to all connected nodes. If new nodes join, or if a node rejoins the cluster after a network partition, then the other nodes in the cluster will replicate all of their registers to the new node.

Caveats

Groot relies on distributed erlang. All of the data stored in Groot is ephemeral and is not maintained between node restarts.

Because we're using CRDTs to propagate changes, a change made on one node may take time to spread to the other nodes. It's safe to run the same operation on multiple nodes. Groot always chooses the register with the latest HLC.

Groot replicates all registers to all nodes. If you attempt to store thousands of keys in Groot, you'll probably have a bad time.

Should I use this?

If you need to store and replicate a relatively small amount of transient values, then Groot may be a good solution for you. If you need anything beyond those features, Groot is probably a bad fit.

Here are some examples of good use cases:

• Feature Flags - rollout is an example of this.
• Runtime configuration changes
• User session state
• Generic caching