Running Akka applications with hundreds and thousands on one system is a scale-up model, where we optimize the usage of the hardware resource available to the application, but within the boundaries of one instance. Soon, the question about how we can scale out our application arises. It explains how we can run our Akka actors on different instances and how they can be managed and controlled as a single entity:
Consider a large grid where various nodes are processing the data and all the worker actors in each of the nodes are managed centrally. Alternatively, think of a map reduce algorithm running where the data is mapped across nodes, which is passed to other nodes for reduction and aggregation:
In the previous diagram, we can see ...