I was wondering the best implementation of a DAG (Directed Acyclic Graph) using Scala/Akka where each node can be potentially on different VMs, different computers. Think of each node as a Scala actor who would only act on its direct parents (in graph terms), how would you implement this at a high level? There are several concepts in Akka that could be applied: • Normal actors with discards if the sender is not a direct parent • Event Bus and Classifiers • … I am still struggling to choose a best approach, at least in theory, knowing that for a DAG, you need to be able to handle diamond scenarios (where a node depends on two or more nodes that in turn depend on one unique node). Any pointers to papers, articles, ideas would be welcomed. Regards
Asked
Active
Viewed 1,822 times