I just playing with the idea.
If I understands well, in an akka cluster every node gets information from every other nodes state-change, and every node can send messages to every other node in that cluster.
What if I want my “not so clustered” app to be more safer, and instead of bind a new node to all of the existing nodes, I want to restrict its information change capability only to x other nodes. So every node would be in a direct contact with x other nodes instead of all the nodes. Every node would communicate only its neighbors, so if some rouge node would try to broadcast some malicious/compute-heavy message, just the neighbors would suffer/stop, and not the whole cluster.
Why this would be good? If I open-source an app, and everybody could host a public node (for example like how most of the block-chains works), one node would can’t kill the whole cluster.
I read the documentation about the data centers and the untrusted mode, but still curious, if some behavior like the upper-one can be implemented with the akka-clusters now (without reinventing the wheel ofc)?