reactive programming in java 8 with rx-java

Post on 07-Jan-2017

11.643 Views

Category:

Software

1 Downloads

Preview:

Click to see full reader

TRANSCRIPT

Reactive Programming in Java 8 with Rx-Java

Kasun IndrasiriSoftware Architect, WSO2

January 2016

Why Reactive Programming? • Why another programming paradigm? • Things are different now! • Conventional software applications won’t be able to fulfill the

modern enterprise needs. • Its time to rethink about our software architecture patterns and

programming techniques.

What it actually means to be Reactive?• “Readily responsive to a stimulus” - merriam-webster

• Reactive Systems :• responds in a timely manner • stays responsive in the face of failure• stays responsive under varying workload• rely on asynchronous message-passing

Reactive Programming• A programming paradigm that helps you to build

‘Reactive Systems’. E.g.: Google Spreadsheet• In a reactive world, we can't just wait for a function

result, a network call, or a database query to return. • Every moment we wait for something, we lose the

opportunity to do other things in parallel. • Reactive programming can be done in several ways• Akka Actors• Reactive Extensions (Rx) – reactive + functional

Why Rx-Java? • The Netflix story..

The Netflix API • How to reducing network chattiness? • Granular API. • client applications to make multiple calls that need to be assembled in order

to render a single user experience

The Netflix API • Discrete requests from devices, should be collapsed into

a single request• Server-side concurrency is needed to effectively reduce

network chattiness.• Nested, conditional, parallel execution of backend

network calls.

Embrace Concurrency• Blocking calls?... Not anymore. • Server-side concurrency is a must. • Low level primitive of concurrency. • Concurrent programming is hard!

Java Futures• Represents the result of an asynchronous computation.• Can be used to retrieve the result asynchronously at a

later point in time• Cons: • Retrieving value from future is a blocking call : Future.get()• Messy with Nested Futures -> complex to implement/maintain.

Callbacks• A callback is a piece of executable code that is passed as

an argument to other code, which is expected to call back the argument at some convenient time.

Callbacks• Use case : • Client send a request to the Application • Application calls service A• Get the response from ServiceA and send it to service B• Get the response from ServiceB and send it to Service C• Get the response from ServiceC and send it to the Client

Service A

Service B

Service C

1

2

3

ApplicationClient

Callbacks• Callback Hell!• Callbacks based applications get extremely complex when we have nested

callbacks. • Leads to messy code, hard to troubleshoot. • Example : https://github.com/kasun04/rnd/tree/master/rx-java/src/main/java/org/kasun/

rnd/rxjava/serviceinvoker/callback

Reactive Extensions - Rx• ReactiveX is a library for composing asynchronous and

event-based programs by using observable sequences.• ‘The Observer pattern done right.’ • Initially implemented for C# by Microsoft, later ported to

many languages by Netflix Rx. • Rx-Java is the Java implementation.

Observer Pattern• ‘Subject’ notifies the ‘observers’ when a change has

happened in the subject.

Observable • Rx-Java introduces ‘Observable’ data type by extending

the traditional Observer Pattern.

Observers

Observable

Observable vs Observer Pattern • The producer signal that there is no more data available:

the• onCompleted() event

• The producer can signal that an error occurred• onError() event

Observable and Subscriber • The producer signal that there is no more data available:

the

ObservableSubscriber

subscribe

onNext*

onCompleted

Observable and Subscriber – In Action

Creating Observable • Observable.create()• Observable.just()• Observable.from()

Observable and Operators • Observable -> Operator -> Subscriber

Filtering Observables• Selectively emit items from a source Observable

Filtering Observables• Filtering Operators

Transforming Observables• Transform items that are emitted by an Observable

Transforming Observables• Transforming Operators

Filtering and Transforming – In Action

Transforming Observables• FlatMap• Observable.flatMap() takes the emissions of one Observable and returns the

emissions of another Observable to take its place.

Observable<List<String>> query(String text){...}

query("Hello, world!") .flatMap(new Func1<List<String>, Observable<String>>() { @Override public Observable<String> call(List<String> urls) { return Observable.from(urls); } }) .subscribe(url -> System.out.println(url));

FlatMap – In Action

More Operators • Combining Observables• Error Handling Operators• Observable Utility Operators• Conditional and Boolean Operators

http://reactivex.io/documentation/operators.html

Rx-Java in Action • Use case : Service Orchestration

Source Code : https://github.com/kasun04/rnd/tree/master/rx-java/src/main/java/org/kasun/rnd/rxjava/serviceinvoker/rx/nonblocking

Service A

Service B

Service C

1

2

3

ApplicationClient

Scheduler• Rx is single threaded by default. • Schedulers are the easiest way to bring multi-threading

into your apps.• subscribeOn• observerOn

Scheduler• Schedulers.io(): meant for I/O-bound work• Schedulers.computation(): meant for computational

work such as event-loops • Schedulers.newThread(): creates a new thread for

each unit of work

Rx-Java in Action • Use case : Service Orchestration with Schedulers

Source Code : https://github.com/kasun04/rnd/tree/master/rx-java/src/main/java/org/kasun/rnd/rxjava/serviceinvoker/rx/nonblocking

Service A

Service B

Service C

1

2

3

ApplicationClient

Conclusion• Get rid of blocking code! • Think ‘non-blocking’ • Think reactive • Rx-Java – steep learning curve • Develop use cases• Leverage Java 8

Questions?

Thank You

top related