@bridget: when we have time we'll need to look at onyx-R and onyx-python 😀

Given a simple topology read -> do-A -> do-B -> write. * .would each node in the topology be considered a peer? * is each peer necessarily a distinct machine? (ec2 instance for example?) * in expanding the cpu capacity of do-A to two machines would each machine represent a peer (effectively expanding the pool from 4 to 5 peers?) In looking at the picture below, I assume their is only ever 1 ring per topology. So the ring both joins together peers with distantly different functions (do-A vs do-B) and those created for scalability (do-A-1) (do-A-2)

A peer is a group of threads that execute a task in your lifecycle. There can be multiple peers per JVM.

In the docs, node=machine=anything running a Jvm

There is one ring per cluster of running peers

Within that ring each peer tries to claim and run jobs