This section provides basic information on the Spring Web Reactive support in Spring Framework 5.
In plain terms reactive programming is about non-blocking applications that are asynchronous and event-driven and require a small number of threads to scale. A key aspect of that definition is the concept of backpressure which is a mechanism to ensure producers don’t overwhelm consumers. For example in a pipeline of reactive components that extends from the database to the HTTP socket when the HTTP client is slow the data repository slows down or stops until capacity frees up.
From a programming model perspective reactive programming involves a major shift from imperative style logic
to a declarative composition of async logic. It is comparable to using CompletableFuture
in Java 8
and composing follow-up actions via lambda expressions.
For a more extended introduction to reactive programming check the excellent multi-part series "Notes on Reactive Programming" by Dave Syer.
Spring Framework 5 embraces
Reactive Streams
as the contract for communicating backpressure across async components and
libraries. Reactive Streams is a specification created through industry collaboration that
has also been adopted in Java 9 as java.util.concurrent.Flow
.
The Spring Framework uses Reactor internally for its own
reactive support. Reactor is a Reactive Streams implementation that further extends the
basic Reactive Streams Publisher
contract with the Flux
and Mono
composable API
types to provide declarative operations on data sequences of 0..N
and 0..1
.
The Spring Framework exposes Flux
and Mono
in many of its own reactive APIs.
At the application level however, as always, Spring provides choice and fully supports
the use of RxJava. For more on reactive types check the post
"Understanding Reactive Types"
by Sebastien Deleuze.
Spring Framework 5 adds a new spring-web-reactive
module that supports the same
@Controller
programming model as Spring MVC but executed on a reactive,
non-blocking engine. The diagram below shows how Spring MVC and Spring Web
Reactive compare side by side:
Spring Web Reactive makes use of Servlet 3.1 non-blocking I/O and runs on
Servlet 3.1 containers. It also runs on non-Servlet runtimes such as Netty and Undertow.
Each runtime is adapted to a set of shared, reactive ServerHttpRequest
and
ServerHttpResponse
abstractions that expose the request and response body
as Flux<DataBuffer>
with full backpressure support on the read and the
write side.
The spring-core
module provides reactive Encoder
and Decoder
contracts
that enable the serialization of a Flux
of bytes to and from typed objects.
The spring-web
module adds JSON (Jackson) and XML (JAXB) implementations for use in
web applications as well as others for SSE streaming and zero-copy file transfer.
The spring-web-reactive
module contains the Spring Web Reactive framework that supports
the @Controller
programming model. It re-defines many of the Spring MVC contracts
such as HandlerMapping
and HandlerAdapter
to be asynchronous and
non-blocking and to operate on the reactive HTTP request and response. For this reason
Spring MVC and Spring Web Reactive cannot share any code. However they do share
many of the same algorithms.
The end result is a programming model identical to today’s Spring MVC but
with support for reactive types and executed in a reactive manner.
For example a controller method can declare a @RequestBody
method argument
in any one of the following ways:
Account account
— the account is deserialized without
blocking before the controller is invoked.
Mono<Account> account
— the controller can use the Mono
to declare logic to be executed after the account is deserialized.
Single<Account> account
— same as with Mono
but using RxJava
Flux<Account> accounts
— input streaming scenario.
Observable<Account> accounts
— input streaming with RxJava.
Similarly a controller can also an @ResponseBody
return value
in any one of the following ways:
Mono<Account>
— serialize without blocking the given Account when the Mono
completes.
Single<Account>
— same but using RxJava.
Flux<Account>
— streaming scenario, possibly SSE depending on the requested content type.
Observable<Account>
— same but using RxJava Observable
type.
Flowable<Account>
— same but using RxJava 2 Flowable
type.
Flux<ServerSentEvent>
— SSE streaming.
Mono<Void>
— request handling completes when the Mono
completes.
void
— request handling completes when the method returns;
implies a synchronous, non-blocking controller method.
Account
— serialize without blocking the given Account;
implies a synchronous, non-blocking controller method.
Spring Framework 5 adds a new reactive WebClient
in addition to the existing RestTemplate
.
Each supported HTTP client (e.g. Reactor Netty) is adapted to a set of shared,
reactive ClientHttpRequest
and ClientHttpResponse
abstractions that expose the request
and response body as Flux<DataBuffer>
with full backpressure support on the read and
the write side. The Encoder
and Decoder
abstractions from spring-core
are also used on
the client side for the serialization of a Flux
of bytes to and from typed objects.
An example of using the WebClient
:
ClientHttpConnector httpConnector = new ReactorClientHttpConnector(); WebClient webClient = new WebClient(httpConnector); Mono<Account> response = webClient .perform(get("http://example.com/accounts/1").accept(APPLICATION_JSON)) .extract(body(Account.class));
The above assumes static method imports from ClientWebRequestBuilders
and ResponseExtractors
that enable a fluent syntax. The same can also be done with RxJava using static imports from
RxJava1ClientWebRequestBuilder
and RxJava1ResponseExtractors
instead:
Single<Account> response = webClient .perform(get("http://example.com/accounts/1").accept(APPLICATION_JSON)) .extract(body(Account.class));
The
Spring Boot Web Reactive starter
available via http://start.spring.io
is the fastest way to get started. It does all that’s necessary so you can start
writing @Controller
classes. By default it runs on Tomcat but the dependencies can
be changed as usual with Spring Boot to switch to a different runtime.
This section outlines the steps to get up and running without Spring Boot.
For dependencies start with spring-web-reactive
and spring-context
.
Then add jackson-databind
and io.netty:netty-buffer:4.1.3.Final
(temporarily see SPR-14528) for JSON support.
Lastly add the dependencies for one of the supported runtimes:
org.apache.tomcat.embed:tomcat-embed-core
org.eclipse.jetty:jetty-server
and org.eclipse.jetty:jetty-servlet
io.projectreactor.ipc:reactor-netty
io.reactivex:rxnetty-common
and io.reactivex:rxnetty-http
io.undertow:undertow-core
For the bootstrap code start with:
ApplicationContext context = new AnnotationConfigApplicationContext(WebReactiveConfiguration.class); // (1) HttpHandler handler = DispatcherHandler.toHttpHandler(context); // (2)
The above loads default Spring Web Reactive config (1), then creates a
DispatcherHandler
, the main class driving request processing (2), and adapts
it to HttpHandler
, the lowest level Spring abstraction for reactive HTTP request handling.
An HttpHandler
can then be installed in each supported runtime:
// Tomcat and Jetty (also see notes below) HttpServlet servlet = new ServletHttpHandlerAdapter(handler); ... // Reactor Netty ReactorHttpHandlerAdapter adapter = new ReactorHttpHandlerAdapter(handler); HttpServer server = HttpServer.create(host, port); server.startAndAwait(adapter); // RxNetty RxNettyHttpHandlerAdapter adapter = new RxNettyHttpHandlerAdapter(handler); HttpServer server = HttpServer.newServer(new InetSocketAddress(host, port)); server.startAndAwait(adapter); // Undertow UndertowHttpHandlerAdapter adapter = new UndertowHttpHandlerAdapter(handler); Undertow server = Undertow.builder().addHttpListener(port, host).setHandler(adapter).build(); server.start();
![]() | Note |
---|---|
For Servlet runtimes you can use the |