Introduction to Reactive Streams and Reactor 2.5

Unless otherwise indicated, these slides are © 2013-2015 Pivotal Software, Inc. and licensed under a Creative Commons Attribution-NonCommercial license: http://creativecommons.org/licenses/by-nc/3.0/
(Rx + Reactive Streams + Spring) % Java 8 =
Reactor 2.5
Stéphane Maldini
@smaldlini

Unless otherwise indicated, these slides are © 2013-2015 Pivotal Software, Inc. and licensed under a 
Creative Commons Attribution-NonCommercial license: http://creativecommons.org/licenses/by-nc/3.0/
Stéphane Maldini
• Survive in London
• Reactive Engineering @
• Project Reactor lead
• Reactive Streams &  
Reactive Streams Commons contributor
• Works on Spring Framework 5 
upcoming Reactive support
• @smaldini on Twitter
2
+

Do you need to be “reactive” ?
If you ask then you probably don’t.
3

How much faster is it ?
It’s slower than your usual imperative system.
4

Creative Commons Attribution-NonCommercial license: http://creativecommons.org/licenses/by-nc/3.0/ 5
Is it good at anything beyond pretending ?
Yes. It scales with the request volume.

Creative Commons Attribution-NonCommercial license: http://creativecommons.org/licenses/by-nc/3.0/ 6
But remember it changes the way you think code.  
It’s definitely a michelin star lunch.

Swipe Right - it’s a match
• Mobile/IoT backend API
• Server-to-Server communications (HTTP…)
• Unreliable clients
• Big Data
• Mutualized Resources (Cloud…)
• User Interfaces
7

Reactive, what the sh*t?
8
More details on http://fr.slideshare.net/StphaneMaldini/intro-to-reactive-programming-52821416
• Reactive is used to broadly define event-driven systems
• Reactive Manifesto defines qualities of reactive systems
• Reactive programming: moving imperative logic to async, non-
blocking, functional-style code, in particular when interacting
with “time consuming” resources

For reactive programming, we need tools :
☐ Reactive Streams
☐ Reactive APIs
9

Reactive Streams
• Reactive Streams is a protocol for asynchronous stream
processing with non-blocking backpressure
• De facto standard for interop between reactive libraries
• To be included in Java 9 as java.util.concurrent.Flow
10

Reactive Streams principle
11
Publisher SubscriberData
Demand
• Max(InflightData) <= demand
• No data sent without demand
• Demand can be unbounded

Reactive Streams is 3+1 interfaces (+ a TCK)
1215
onSubscribe
onNext*
(onError|onComplete)?
Publisher
Subscriber
Subscription
request(n)
cancel()

Reactive Streams offers Quality of Service for the JVM
Bounded hardware use and flow prioritization
13

Imperative vs Reactive Streams ?
14
User rickSanchez =
userRepository.findUser(“rick”);
Blocking, not
returning until done !
Might throw
exceptions

If this blocks, and runs on a serving HTTP request ?
If the calling HTTP requests come from another “Microservice” ?
If that Microservice is also calling in a blocking way ?
If you solve scalability issues by scaling out ?
If you agreed to pay the scalability tax, how far can it help you ?
Thinking about the big picture
15

Imperative vs Reactive Streams ?
16
Publisher<User> rickSanchez =
rickSanchez.subscribe(new Subscriber<User>(){ … });
Might send 0, 1 or N Users ! Non Blocking - on demand data emission
Callback for start, result, error or complete

How practical is Publisher ?
17

Do you want to implement Publisher yourself ?
Probably not, usually.
18

For reactive programming, we need tools :
☑ Reactive Streams
☐ Reactive APIs
19

• Transform, Combine, Timebox, Aggregate, Consume…
• On the JVM:
• Reactor 2.5 is 4th generation* and based on Reactive Streams
• RxJava 1.x: 2nd generation* and most used implementation
• Akka Stream 2.x: Lightbend 3rd generation* Reactive API
• Also for other languages, for example RxJS, MostJS
Reactive APIs
20
* Based on http://akarnokd.blogspot.fr/2016/03/operator-fusion-part-1.html

Reactor 2.5 ecosystem
21

Reactor Core 2.5
• Cross collaboration with Dávid Karnok (RxJava) and some
Spring Framework committers
• Natively based on Reactive Streams, RSC* and Java 8+
• 2 new rich Publisher types : Flux & Mono
• Strong focus on efficiency
• Ever-improving debugging, logging, testing capabilities
* ReactiveStreamsCommons is a research effort about reactive flows
22

Flux (0..N elements) with ReactiveX compliant API
23

Mono (0..1 element)
24

Imperative vs Reactor ?
25
Mono<User> rickSanchez =
rickSanchez.consume(this::userHandler,
this::errorHandler)
Single Typed Publisher !
Start the Mono and consume
its result or error.

Mono, a single-data-at-most API
26

Flux, classic Rx patterns for 0..N events
27

Visual (Marble) doc for Flux & Mono
28

Reactor Web Console (preview)
29

Concurrency Types Cheat Sheet
30
No value Single value Multiple values
Blocking void T 
Future<T>
Iterable<T>
Collection<T>
Stream<T>
Non-
blocking
CompletableFuture<Void> CompletableFuture<T> CompletableFuture<List<T>>
Reactive 
Streams
Publisher<Void> Publisher<T> Publisher<T>
RxJava Observable<Void>
Completable (1.1.1)
Observable<T>
Single<T> (1.0.13)
Observable<T>
Reactor Mono<Void> Mono<T> Flux<T>

https://spring.io/blog/2016/04/19/understanding-reactive-types
31

Reactive Spring
32
• Spring projects are going reactive
• Reactor Core is the reactive foundation
• RxJava support provided out of the box
• You will be able to choose your web engine: 
Tomcat, Jetty, Undertow or Netty
• Most impact on Web and Data support (IO intensive)
• Spring Reactive experiment
• Spring Reactive Playground sample application

Well known Controller example
33
@RestController
public class UserController {
private BlockingRepository<User> repository;
@RequestMapping(path = "/save-capitalized", method = RequestMethod.POST)
public void saveCapitalized(@RequestBody List<User> users) {
users.forEach(u -> u.setName(u.getName().toUpperCase()));
repository.save(users);
}
}
public interface BlockingRepository<T> {
void save(List<T> elements);
Iterable<T> findAll();
}

Controller with Reactive types
34
@RestController
private ReactiveRepository<User> repository;
@RequestMapping(path = "/save-capitalized", method = RequestMethod.POST)
public Mono<Void> saveCapitalized(@RequestBody Flux<User> users) {
return repository.save(users.map(u -> new User(u.getName().toUpperCase()));
}
}
public interface ReactiveRepository<T> {
Mono<Void> save(Publisher<T> elements);
Flux<T> findAll();
}

Blocking vs Reactive: memory consumption
35
Memoryconsumption
Time
Blocking Reactive

Blocking vs Reactive: streaming updates
36
Numberofuserssaved
inthedatabase
Time
Blocking Reactive

Controller with Reactive return values
37
@RestController
private ReactiveRepository<User> repository;
@RequestMapping(path = "/", method = RequestMethod.GET)
public Flux<User> findAll() {
return repository.findAll();
}
}
• Optimized serialization when using Flux instead of List
• Also perfectly suitable for Server-Sent Events

Reactive HTTP client with Mono
38
import static org.springframework.web.client.reactive.HttpRequestBuilders.*;
import static org.springframework.web.client.reactive.WebResponseExtractors.*;
Mono<Person> result = webClient
.perform(get("http://localhost:8080/person")
.header("X-Test-Header", "testvalue")
.accept(MediaType.APPLICATION_JSON))
.extract(body(Person.class));

Reactive HTTP client with Flux
39
import static org.springframework.web.client.reactive.HttpRequestBuilders.*;
import static org.springframework.web.client.reactive.WebResponseExtractors.*;
Flux<Person> response = webClient
.perform(get("http://localhost:8080/persons")
.accept(MediaType.APPLICATION_JSON))
.extract(bodyStream(Person.class));
Works for:
• JSON array [{"foo":"bar"},{"foo":"baz"}]
• JSON Streaming {"foo":"bar"}{"foo":"baz"}
• SSE with something like .extract(sseStream(Person.class))

Do you want to learn and play at the same time ?
Get Factorio on Steam for a few coins
40

Introduction to Reactive Streams and Reactor 2.5

Empfohlen

Empfohlen

Weitere ähnliche Inhalte

Was ist angesagt?

Was ist angesagt? (20)

Andere mochten auch

Andere mochten auch (8)

Ähnlich wie Introduction to Reactive Streams and Reactor 2.5

Ähnlich wie Introduction to Reactive Streams and Reactor 2.5 (20)

Mehr von Stéphane Maldini

Mehr von Stéphane Maldini (13)

Kürzlich hochgeladen

Kürzlich hochgeladen (20)

Introduction to Reactive Streams and Reactor 2.5