Akka threads running on cluster nodes

Question

I am trying to understand the large-scale part of Akka threads. Here I see various answers, but still do not understand how this functionality is implemented.

There are ActorPublisher / Subscribers, which, from what I understand, are designed to interact with the stream, but at the same time they say that they cannot be deleted.

How could I use the ability to use akka toegether clusters with threads? Should I outsource GraphStage to my remote player? Or can graphStage itself run in a node cluster?

In my use case, there is a pipeline containing several heavy calculation steps with merges and merges, and maybe some of these steps should be distributed.

Update
According to this Internet Bypass with Akka Streams, I see that a single graph cannot span multiple clusters. What will be the horizontal scale in this case?

Answer 1

akka-streams does not yet support clustering. I suggest using Apache Kafka as a middleware for transport, if you need some clustering solution, and reactive-kafka . But, of course, you can implement your own transport layer with actors and run it in the Akka cluster.

Answer 2

This may be a late reply. But there is a good article that sums up some integration between akka threads and the cluster. https://blog.colinbreck.com/integrating-akka-streams-and-akka-actors-part-iii/