Revitalizing Enterprise Integration with Reactive Streams
1 like5,430 views
The document discusses the potential of reactive enterprise integration using reactive streams, emphasizing the importance of streaming as a core concept for improving flow control in data processing. It highlights the advantages of non-blocking architectures and provides an overview of the capabilities of reactive streams in Java and Scala, including built-in flow control and support for varying data rates. The talk references various connectors available in Alpakka, underscoring the tools and resources available for implementing these integration patterns.
Revitalizing Enterprise Integration with Reactive Streams
- 1. Revitalizing Enterprise Integration with Reactive Streams √ Deputy CTO at Lightbend @viktorklang / [email protected] / http://viktorklang.com
- 3. What to expect from this talk?
- 4. The most common programming tasks? 4 Naming things 4 Design the Perfect Bikeshed 4 Receive inputs 4 Transform data 4 Produce outputs
- 5. code.demo()
- 6. Systems
- 7. Step 1: connect(ALL THE THINGS) Step 2: Ship it!
- 8. ||
- 9. The past and present of Enterprise Integration
- 10. Now playing: «Why SOA failed (Klang remix)» 4 WS 4 Schema evolution 4 Service Discovery 4 Testing 4 ESBs
- 11. The rise of Enterprise Integration Patterns1 1 Gregor Hohpe and Bobby Woolf, 2003
- 12. Breaking the Camel's Back?
- 13. What if…
- 14. Reactive Enterprise Integration? 4 Streaming as a first-class concept 4 Flow control built-in 4 Reusable shapes for integration
- 15. Flow Control Facts 4 Buffers hide latency 4 Buffers do not solve overload problems 4 Load shedding does not inform sender 4 Blocking creates a low upper limit on concurrency 4 Blocking makes fan-in and fan-out problematic
- 16. Flow Control Done Right 4 Support potentially unbounded sequences 4 Sender runs independently from receiver 4 Rate of reception may vary from rate of sending 4 Dropping data should be a choice and not a necessity 4 Minimal—if any—overhead in terms of latency and throughput
- 17. Reactive Streams2 & java.util.concurrent.Flow 4 Publisher<Output> 4 Subscriber<Input> 4 Processor<Input, Output> 4 Subscription 2 www.reactive-streams.org
- 18. Reactive Streams
- 19. Reactive Streams & java.util.concurrent.Flow 4 «push» when subscriber is faster 4 «pull» when publisher is faster 4 switches dynamically between both 4 smart prefetch made possible
- 20. 4 immutable 4 reusable 4 composable 4 coordinated 4 asynchronous 4 distributed 4 transformations
- 21. Akka Stream building blocks3 3 Non-exhaustive list
- 23. Riding the Alpakka4 Reactive Enterprise Integration based on Akka Streams—for Java and Scala. 4 https://github.com/akka/alpakka (Apache v2)
- 24. code.demo()
- 25. Connectors currently found in Alpakka5 AMQP, Apache Geode, Apache Solr, AWS DynamoDB, AWS Kinesis, AWS Lambda, AWS S3, AWS SNS, AWS SQS, Azure Storage Queue, Cassandra, Elasticsearch, File IO, FTP, Google Cloud Pub/Sub, HBase, IronMq, JMS, MongoDB, MQTT, OrientDB, Server-sent Events, Slick (JDBC), Spring Web, Unix Domain Socket 5 And also HTTP, TCP, Apache Kafka, and many more…
- 26. Further reading Transforming enterprise integration with reactive streams https://www.oreilly.com/ideas/transforming-enterprise-integration-with-reactive-streams Reactive Programming vs Reactive Systems https://www.oreilly.com/ideas/reactive-programming-vs-reactive-systems
- 27. References 4 The example: https://github.com/lightbend/ reactive-integration-examples/tree/master/ integration-webinar 4 Akka Streams documentation: https://doc.akka.io/ docs/akka/current/stream/index.html? language=java 4 Alpakka documentation: https:// developer.lightbend.com/docs/alpakka/current/
- 28. Thank you!