carrot-garden / akka_akka-data-replication

Replication of CRDTs in Akka Cluster

Akka Data Replication

This is an EARLY PREVIEW of a library for replication of data in an Akka cluster. It is a replicated in-memory data store supporting low latency and high availability requirements. The data must be so called Conflict Free Replicated Data Types (CRDTs), i.e. they provide a monotonic merge function and the state changes always converge.

For good introduction to CRDTs you should watch the Eventually Consistent Data Structures talk by Sean Cribbs.

CRDTs can't be used for all types of problems, but when they can they have very nice properties:

• low latency of both read and writes
• high availability (partition tolerance)
• scalable (no central coordinator)
• strong eventual consistency (eventual consistency without conflicts)

Built in data types:

• Counters: GCounter, PNCounter
• Registers: LWWRegister, Flag
• Sets: GSet, ORSet
• Maps: ORMap, LWWMap, PNCounterMap

You can use your own custom data types by implementing the merge function of the ReplicatedData trait. Note that CRDTs typically compose nicely, i.e. you can use the provided data types to build richer data structures.

The Replicator actor implements the infrastructure for replication of the data. It uses direct replication and gossip based dissemination. The Replicator actor is started on each node in the cluster, or group of nodes tagged with a specific role. It communicates with other Replicator instances with the same path (without address) that are running on other nodes. For convenience it is typically used with the DataReplication Akka extension.

A short example of how to use it:

class DataBot extends Actor with ActorLogging {
  import DataBot._
  import Replicator._

  val replicator = DataReplication(context.system).replicator
  implicit val cluster = Cluster(context.system)

  import context.dispatcher
  val tickTask = context.system.scheduler.schedule(5.seconds, 5.seconds, self, Tick)

  replicator ! Subscribe("key", self)

  def receive = {
    case Tick =>
      val s = ThreadLocalRandom.current().nextInt(97, 123).toChar.toString
      if (ThreadLocalRandom.current().nextBoolean()) {
        // add
        log.info("Adding: {}", s)
        replicator ! Update("key", ORSet(), WriteLocal)(_ + s)
      } else {
        // remove
        log.info("Removing: {}", s)
        replicator ! Update("key", ORSet(), WriteLocal)(_ - s)
      }

    case _: UpdateResponse => // ignore

    case Changed("key", data: ORSet) =>
      log.info("Current elements: {}", data.value)
  }

  override def postStop(): Unit = tickTask.cancel()

}

The full source code for this sample is in DataBot.scala.

More detailed documentation can be found in the ScalaDoc of Replicator and linked classes.

Two other examples:

• VotingService
• ShoppingCart

Dependency

Latest version of akka-data-replication is 0.8. This version depends on Akka 2.3.6 and is cross-built against Scala 2.10.4 and 2.11.4.

Add the following lines to your build.sbt file:

resolvers += "patriknw at bintray" at "http://dl.bintray.com/patriknw/maven"

libraryDependencies += "com.github.patriknw" %% "akka-data-replication" % "0.8"

More Resources

• Eventually Consistent Data Structures talk by Sean Cribbs
• Strong Eventual Consistency and Conflict-free Replicated Data Types talk by Mark Shapiro
• A comprehensive study of Convergent and Commutative Replicated Data Types paper by Mark Shapiro et. al.