Awesome AKKA

Author: Homer Quan (hi@homerquan.me)

SCALA and AKKA

• How to code in Scala
  • (A quick intro from twitter) https://twitter.github.io/scala_school/basics.html
  • Effective scala: http://twitter.github.io/effectivescala/
  • https://class.coursera.org/progfun-004
  • https://www.coursera.org/curse/reactive
  • (A style guide) http://docs.scala-lang.org/style/
  • (ref) http://www.scala-lang.org/docu/files/ScalaReference.pdf
• Become an expert
  • http://danielwestheide.com/scala/neophytes.html
  • http://www.slideshare.net/shinolajla/effective-akka-scalaio
  • using typesafe console http://resources.typesafe.com/docs/console/manual/getting-started.html
  • https://github.com/alexandru/scala-best-practices
  • Some useful libs for scala: https://github.com/lauris/awesome-scala
• scala projects
  • Official samples: https://github.com/akka/akka-samples
  • github https://github.com/trending?l=scala
  • template http://www.typesafe.com/activator/templates
  • alternative template: https://github.com/dph01/scala-sbt-template
• Start your first scala project: http://www.johnmurray.io/log/2014/04/15/Starting-a-Scala-Project.html
• Best practices in coding: https://github.com/alexandru/scala-best-practices

Why Scala

JVM and Java compatibility Functional and OO style: productivity Actor: flexible scale up/down, maintainable, fault tolerance * Real time: Scala is built for stream, async processing (lambda, future, promise)

Why AKKA

https://www.typesafe.com/community/core-projects/akka http://blog.confluent.io/2015/01/29/making-sense-of-stream-processing/ http://www.typesafe.com/blog/reactive-streams-webinar-qa

How to code in Akka stream

http://doc.akka.io/docs/akka-stream-and-http-experimental/1.0-M2/scala/stream-cookbook.html

Async style programming

Scala and AKKA is non-blocking async language. Please familiar yourself with concepts such as Future and Promise

Actors

Scala and AKKA provides a concurrency mechanism called “Actor”. It would benefit WISE in scalability and resilience.

Actor is an isolated resource unit and communicate with each other via messages. It needs a storage persistence to recover from failure or bootstrap. We select cassandra as storage provider. Sample code: https://github.com/krasserm/akka-persistence-cassandra

Find vehicles by spitial query: build index in memory http://jspatial.sourceforge.net/

Actor is built for share nothing. But sometime it has to use share status: http://stackoverflow.com/questions/7307217/akka-sharing-mutable-state

Common Actor patterns https://doc.akka.io/docs/akka/2.5.4/scala/howto.html

HA

Akka persistent: http://www.slideshare.net/ktoso/akka-persistence-event-sourcing-in-30-minutes

Multi tenant

• Each account will has its own actor system

AKKA streem

• Akka stream cookbook http://doc.akka.io/docs/akka-stream-and-http-experimental/1.0-M2/scala/stream-cookbook.html
• Stream graph http://doc.akka.io/docs/akka-stream-and-http-experimental/1.0-M4/scala/stream-graphs.html
• Akka stream custom process stage (writing nodes in CEP network) http://doc.akka.io/docs/akka-stream-and-http-experimental/1.0-M4/scala/stream-customize.html
• Using actor with stream: http://doc.akka.io/docs/akka-stream-and-http-experimental/1.0-M4/scala/stream-integrations.html
• http://doc.akka.io/docs/akka-stream-and-http-experimental/1.0-M5/stream-design.html
• Integrating Akka Streams and Actors http://blog.colinbreck.com/integrating-akka-streams-and-akka-actors-part-i/

Important cencepts

• Graph: Flow (lines in the graph e.g., f1, f2 … ) + source (in) + sink (out) + junctions (bcast and merge)
• Flow: A linear processing unit composited by stages e.g., f1: s1=>s2=>s3
• Stage: The basic processing unit, which can be
  • pull
  • push
  • pushpull (for most stream processing)
  • stateful (if change internal status)

Code samples

• Processing node https://github.com/carlpulley/lift/blob/model/server/exercise/src/main/scala/com/eigengo/lift/exercise/classifiers/workflows/SlidingWindow.scala
• Unit test: https://github.com/carlpulley/lift/blob/model/server/exercise/src/test/scala/com/eigengo/lift/exercise/classifiers/workflows/SlidingWindowTest.scala

Actors

• Actors for resilliance and scalability http://doc.akka.io/docs/akka/snapshot/scala/persistence.html, http://www.slideshare.net/bantonsson/akka-persistencescaladays2014
• Message dispatcher in AKKA actor system http://doc.akka.io/docs/akka/snapshot/scala/dispatchers.html
• An akka actor example (stock) https://github.com/typesafehub/reactive-stocks#master
• Use actor with sream http://doc.akka.io/docs/akka-stream-and-http-experimental/1.0-M3/scala/stream-integrations.html http://doc.akka.io/api/akka-stream-and-http-experimental/0.7/index.html#akka.stream.actor.ActorPublisher
• http://www.slideshare.net/NLJUG/akka-in-practice-age-mooij-and-raymond-roestenburg
• http://stackoverflow.com/questions/3931994/design-patterns-best-practice-for-building-actor-based-system
• Integration http://doc.akka.io/docs/akka/snapshot/scala/camel.html
• Actor pattern http://letitcrash.com/post/30585282971/discovering-message-flows-in-actor-systems-with
• EPI pattern https://gist.github.com/jboner/8556813
• HA http://danielwestheide.com/blog/2013/03/20/the-neophytes-guide-to-scala-part-15-dealing-with-failure-in-actor-systems.html

Other

• FSM in AKKA http://doc.akka.io/docs/akka/snapshot/scala/fsm.html
• AKKA CEP with stream http://typesafe.com/activator/template/akka-with-esper#code/src/main/scala/experiments/streams/StreamCEP.scala
• Scala scheduler http://doc.akka.io/docs/akka/2.0/scala/scheduler.html
• Scala agent http://doc.akka.io/docs/akka/2.1.0/scala/agents.html
• A CEP based on spark stream http://spark-summit.org/2014/talk/stratio-streaming-a-new-approach-to-spark-streaming
• http://doc.akka.io/docs/akka/2.0/scala/dataflow.html
• http://www.scala-lang.org/api/2.10.4/index.html#scala.util.continuations.package
• https://github.com/scala/async
• Streaming and online algorithms for Akka: https://github.com/gosubpl/akka-online