Spring Reactive Programming

6 years ago by Anil

Introduction to Reactive Programming

Reactive Programming is a program pattern for an asynchronous, non-blocking, event-driven approach for data processing.
Data processing based on events triggered by changes in streams.
In Reactive Programming small number of threads are enough for processing data.
In Reactive Programming everything is Streams, it reacts based on changes of data.

Why Reactive Programming

Reactive Programming is more focus on independent of Thread framework, so code is more readable.
Everything is event-based, so programmers focus on mostly on business logic.
Reactive Programming is the best control between producer and consumer with backpressure control.
Highly interactive for user interface also.

Blocking HTTP requests

In web applications, a web server is capable of a limited number of HTTP requests, Tomcat server limited to 50 requests by default.
For each and every request server will create separate Thread, most web server creates 50 threads based on server requests limit.
If we get 51th request server will not process and waits for existing threads free.
All new requests will be in blocking called as synchronous blocking request processing.

Non-Blocking HTTP requests

In non-blocking asynchronous HTTP requests, no threads in waiting state.
When a request comes it process by a separate thread but it is reactive nature it triggers event for IO response or DB response.
In this model server can handle more number of requests.

Reactive Streams

Reactive Streams API created by Ngnix, Pivotal, Oracle team engineers initially.
Later on its a part of JDK - 9 with major 4 interfaces.

Publisher
Subscriber
Subscription
Processor

Publisher

public interface Publisher<T>
{
    public void subscribe(Subscriber<? super T> s);
}

Subscriber

public interface Subscriber<T>
{
    public void onSubscribe(Subscription s);
    public void onNext(T t);
    public void onError(Throwable t);
    public void onComplete();
}

Subscriber

public interface Subscription<T>
{
    public void request(long n);
    public void cancel();
}

Subscriber

public interface Processor<T, R> extends Subscriber<T>, Publisher<R>
{
}

Spring WebFlux

Spring WebFlux module supports non-blocking asynchronous data processing.
It supports backpressure by default.
By using netty server internally to run the reactive application.
Spring webflux uses project reactor as reactive library.
Spring WebFlux has 2 publisher interfaces

Mono
Flux