Common algos framework for erlang (Caffe)

• Is a framework to create and deploy distributed algos on a graph G = (V, E).
• Separates user logic from distributed algos running on each vertex
• Distributed algos are composable via plug-ins

Example Usage

G = {V, E} = caffe_graph:load(graph1),
Network = caffe:build( G, caffe:args_from_list( [ { V, { [ lamport_clock ], worker_random_messenger } } ] ) ),
caffe:start( Network )

• We load a graph from a module using caffe_graph:load/1
• worker_random_messenger represents the user-based function running on each vertex in V
  This function sends a unique message to a random outgoing edge.
• [ lamport_clock ] represents the set of plugins/distributed algorithms we want to run on top of the user-based function
  This algorithm creates a logical clock based on send and receive events
  You may specify multiple plugins, and plugins that use other plugins (dependencies) are auto-loaded