Reactive programming with Rxjava
2 likes1,227 views
The document discusses reactive programming using RxJava, highlighting the challenges and benefits of asynchronous and non-blocking system architecture. It emphasizes the importance of efficient resource utilization and the need for parallelizing code due to limitations imposed by Amdahl's law. The presentation also includes references to various resources and examples related to RxJava applications and operators.
![https://github.com/toff63/Sandbox/blob/master/java/rsjug-rx/rsjug-rx/src/main/java/rs/jug/rx/flowcontrol/Backpressure.java
Buffer
[0, 1, 2, 3, 4, 5]
[6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20]
[21, 22, 23, 24, 25]](https://image.slidesharecdn.com/reactiveprogrammingwithrxjava-161005012327/85/Reactive-programming-with-Rxjava-57-320.jpg)
![Stacktrace example
18:42:59.487 [rx-request-processor-5-67] ERROR
n.k.t.util.HttpContentInputStream - Error on server
io.netty.util.IllegalReferenceCountException: refCnt: 0
at io.netty.buffer.AbstractByteBuf.ensureAccessible(AbstractByteBuf.java:1178) ~[netty-buffer-4.0.27.Final.jar:4.0.27.Final]
at io.netty.buffer.AbstractByteBuf.writeBytes(AbstractByteBuf.java:848) ~[netty-buffer-4.0.27.Final.jar:4.0.27.Final]
at io.netty.buffer.AbstractByteBuf.writeBytes(AbstractByteBuf.java:841) ~[netty-buffer-4.0.27.Final.jar:4.0.27.Final]
at io.netty.buffer.AbstractByteBuf.writeBytes(AbstractByteBuf.java:831) ~[netty-buffer-4.0.27.Final.jar:4.0.27.Final]
at netflix.karyon.transport.util.HttpContentInputStream$1.onNext(HttpContentInputStream.java:67) [karyon2-governator-2.3.0.jar:2.3.0]
at netflix.karyon.transport.util.HttpContentInputStream$1.onNext(HttpContentInputStream.java:33) [karyon2-governator-2.3.0.jar:2.3.0]
at rx.Observable$33.onNext(Observable.java:7480) [rxjava-1.0.10.jar:1.0.10]
at rx.observers.SafeSubscriber.onNext(SafeSubscriber.java:130) [rxjava-1.0.10.jar:1.0.10]
at io.reactivex.netty.protocol.http.UnicastContentSubject$AutoReleaseByteBufOperator$1.onNext(UnicastContentSubject.java:262) [rxnetty-0.4.9.jar:0.4.9]
at rx.internal.operators.NotificationLite.accept(NotificationLite.java:150) [rxjava-1.0.10.jar:1.0.10]
at rx.internal.operators.BufferUntilSubscriber.emit(BufferUntilSubscriber.java:151) [rxjava-1.0.10.jar:1.0.10]
at rx.internal.operators.BufferUntilSubscriber.onNext(BufferUntilSubscriber.java:184) [rxjava-1.0.10.jar:1.0.10]
at io.reactivex.netty.protocol.http.UnicastContentSubject.onNext(UnicastContentSubject.java:286) [rxnetty-0.4.9.jar:0.4.9]
at io.reactivex.netty.protocol.http.server.ServerRequestResponseConverter.invokeContentOnNext(ServerRequestResponseConverter.java:193)
[rxnetty-0.4.9.jar:0.4.9]
at io.reactivex.netty.protocol.http.server.ServerRequestResponseConverter.channelRead(ServerRequestResponseConverter.java:129) [rxnetty-0.4.9.jar:0.4.9]
at io.netty.channel.AbstractChannelHandlerContext.invokeChannelRead(AbstractChannelHandlerContext.java:339) [netty-transport-4.0.27.Final.jar:4.0.27.Final]
at io.netty.channel.AbstractChannelHandlerContext.access$600(AbstractChannelHandlerContext.java:32) [netty-transport-4.0.27.Final.jar:4.0.27.Final]
at io.netty.channel.AbstractChannelHandlerContext$7.run(AbstractChannelHandlerContext.java:329) [netty-transport-4.0.27.Final.jar:4.0.27.Final]
at io.netty.util.concurrent.DefaultEventExecutor.run(DefaultEventExecutor.java:36) [netty-common-4.0.27.Final.jar:4.0.27.Final]
at io.netty.util.concurrent.SingleThreadEventExecutor$2.run(SingleThreadEventExecutor.java:111) [netty-common-4.0.27.Final.jar:4.0.27.Final]
at io.netty.util.concurrent.DefaultThreadFactory$DefaultRunnableDecorator.run(DefaultThreadFactory.java:137) [netty-common-4.0.27.Final.jar:4.0.27.Final]
at java.lang.Thread.run(Thread.java:745) [na:1.8.0_73]](https://image.slidesharecdn.com/reactiveprogrammingwithrxjava-161005012327/85/Reactive-programming-with-Rxjava-73-320.jpg)
Reactive programming with Rxjava
- 1. Christophe Marchal Reactive programming with RxJava
- 3. MOTIVATION
- 4. Scaling Moore’s law is delivering more cores but not faster cores
- 5. Amdahl’s law Speedup limited by the sequential portion of the code
- 6. In other words, parallelize your code to scale
- 7. I am not Netflix!
- 9. Efficient use of resources ● CPU efficient ● Memory efficient ● Hard Drive efficient ● Network efficient
- 10. Efficient use of threads
- 12. http://techblog.netflix.com/2016/09/zuul-2-netflix-journey-to-asynchronous.html Asynchronous and Non-Blocking System Architecture
- 13. Challenges ● listeners / callback ● force functional code ● exception handling ● Everything becomes a Stream
- 14. RxJava for the win!
- 15. Challenges
- 19. Observable: Stream of event Reactive vision Everything is an event
- 20. Observable: Stream of event Observer Reactive vision
- 21. Reactive vision T getData() One Item synchronous
- 22. Reactive vision Iterable<T> getData()T getData() One Item Several Items synchronous
- 23. Reactive vision Iterable<T> getData()T getData() Future<T> getData() One Item Several Items synchronous Asynchronous
- 24. Reactive vision Iterable<T> getData() Observable<T> getData() T getData() Future<T> getData() One Item Several Items synchronous Asynchronous
- 25. Pull vs Push Iterable<T> getData() Observable<T> getData() Pull T next() throw Exception() returns; Push onNext() onError() onComplete()
- 35. https://raw.github.com/wiki/ReactiveX/RxJava/images/rx-operators/zip.png Zip
- 39. Demo
- 40. Concurrency
- 41. Observable is sequential Scheduling and Combining Observables enable Concurrency
- 43. Demo
- 44. Cold Stream vs Hot Stream Hot Stream Cold Stream no control on emission rate emits when requested UI events, Metric events, System events DB query, Service request, Downloading file
- 45. Cold Stream vs Hot Stream Hot Stream Cold Stream no control on emission rate emits when requested UI events, Metric events, System events DB query, Service request, Downloading file Flow control Flow control & Back pressure
- 46. Flow Control
- 49. Block Operator Hot Stream Cold Stream
- 54. Debounce
- 56. Buffer
- 57. https://github.com/toff63/Sandbox/blob/master/java/rsjug-rx/rsjug-rx/src/main/java/rs/jug/rx/flowcontrol/Backpressure.java Buffer [0, 1, 2, 3, 4, 5] [6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20] [21, 22, 23, 24, 25]
- 58. Window time
- 60. Reactive Stream
- 61. ● Push when Consumer keeps up ● Pull when Consumer is slow ● Bound all queues Maximize throughput
- 62. Reactive Stream: Consumer keeps up Publisher Subscriber As many as you can Push
- 63. Reactive Stream: Consumer Starts buffering Publisher Subscriber As many as you can
- 64. Reactive Stream: Consumer Starts buffering Publisher Subscriber Give me 0 Give me 2 Pull
- 66. On backpressure buffer Hot Stream Hot Stream
- 67. On backpressure buffer Scheduler Hot Stream
- 69. On backpressure buffer Hot Stream
- 70. And lots and lots of other operators
- 71. Rx Ports
- 72. ● Observable API is complex ● Takes time to become fluent with Observable ● Hard to test ! ● Debugging is harder as everything is asynchronous ● Stacktraces can be truncated due to scheduler Drawbacks
- 73. Stacktrace example 18:42:59.487 [rx-request-processor-5-67] ERROR n.k.t.util.HttpContentInputStream - Error on server io.netty.util.IllegalReferenceCountException: refCnt: 0 at io.netty.buffer.AbstractByteBuf.ensureAccessible(AbstractByteBuf.java:1178) ~[netty-buffer-4.0.27.Final.jar:4.0.27.Final] at io.netty.buffer.AbstractByteBuf.writeBytes(AbstractByteBuf.java:848) ~[netty-buffer-4.0.27.Final.jar:4.0.27.Final] at io.netty.buffer.AbstractByteBuf.writeBytes(AbstractByteBuf.java:841) ~[netty-buffer-4.0.27.Final.jar:4.0.27.Final] at io.netty.buffer.AbstractByteBuf.writeBytes(AbstractByteBuf.java:831) ~[netty-buffer-4.0.27.Final.jar:4.0.27.Final] at netflix.karyon.transport.util.HttpContentInputStream$1.onNext(HttpContentInputStream.java:67) [karyon2-governator-2.3.0.jar:2.3.0] at netflix.karyon.transport.util.HttpContentInputStream$1.onNext(HttpContentInputStream.java:33) [karyon2-governator-2.3.0.jar:2.3.0] at rx.Observable$33.onNext(Observable.java:7480) [rxjava-1.0.10.jar:1.0.10] at rx.observers.SafeSubscriber.onNext(SafeSubscriber.java:130) [rxjava-1.0.10.jar:1.0.10] at io.reactivex.netty.protocol.http.UnicastContentSubject$AutoReleaseByteBufOperator$1.onNext(UnicastContentSubject.java:262) [rxnetty-0.4.9.jar:0.4.9] at rx.internal.operators.NotificationLite.accept(NotificationLite.java:150) [rxjava-1.0.10.jar:1.0.10] at rx.internal.operators.BufferUntilSubscriber.emit(BufferUntilSubscriber.java:151) [rxjava-1.0.10.jar:1.0.10] at rx.internal.operators.BufferUntilSubscriber.onNext(BufferUntilSubscriber.java:184) [rxjava-1.0.10.jar:1.0.10] at io.reactivex.netty.protocol.http.UnicastContentSubject.onNext(UnicastContentSubject.java:286) [rxnetty-0.4.9.jar:0.4.9] at io.reactivex.netty.protocol.http.server.ServerRequestResponseConverter.invokeContentOnNext(ServerRequestResponseConverter.java:193) [rxnetty-0.4.9.jar:0.4.9] at io.reactivex.netty.protocol.http.server.ServerRequestResponseConverter.channelRead(ServerRequestResponseConverter.java:129) [rxnetty-0.4.9.jar:0.4.9] at io.netty.channel.AbstractChannelHandlerContext.invokeChannelRead(AbstractChannelHandlerContext.java:339) [netty-transport-4.0.27.Final.jar:4.0.27.Final] at io.netty.channel.AbstractChannelHandlerContext.access$600(AbstractChannelHandlerContext.java:32) [netty-transport-4.0.27.Final.jar:4.0.27.Final] at io.netty.channel.AbstractChannelHandlerContext$7.run(AbstractChannelHandlerContext.java:329) [netty-transport-4.0.27.Final.jar:4.0.27.Final] at io.netty.util.concurrent.DefaultEventExecutor.run(DefaultEventExecutor.java:36) [netty-common-4.0.27.Final.jar:4.0.27.Final] at io.netty.util.concurrent.SingleThreadEventExecutor$2.run(SingleThreadEventExecutor.java:111) [netty-common-4.0.27.Final.jar:4.0.27.Final] at io.netty.util.concurrent.DefaultThreadFactory$DefaultRunnableDecorator.run(DefaultThreadFactory.java:137) [netty-common-4.0.27.Final.jar:4.0.27.Final] at java.lang.Thread.run(Thread.java:745) [na:1.8.0_73]
- 74. Netflix experience migrating zuul http://techblog.netflix.com/2016/09/zuul-2-netflix-journey-to-asynchronous.html
- 75. References ● https://www.infoq.com/presentations/rx-service-archite cture ● https://www.infoq.com/presentations/rxjava-reactor ● http://reactivex.io/tutorials.html ● http://reactivex.io/documentation/operators.html
- 76. Christophe Marchal Thank you ! Questions?