Table of Contents
As Guacamole is an API, one of the best ways to put Guacamole to use is by building your own Guacamole-driven web application, integrating HTML5 remote desktop into whatever you think needs it.
The Guacamole project provides an example of doing this called "guacamole-example", but this example is already completed for you, and from a quick glance at this example, it may not be obvious just how easy it is to integrate remote access into a web application. This tutorial will walk you through the basic steps of building an HTML5 remote desktop application using the Guacamole API and Maven.
Guacamole's architecture is made up of many components, but it's actually straightforward, especially from the perspective of the web application.
Guacamole has a proxy daemon, guacd, which handles communication using remote desktop protocols, exposing those to whatever connects to it (in this case, the web application) using the Guacamole protocol. From where the web application is standing, it doesn't really matter that guacd dynamically loads protocol plugins or that it shares a common library allowing this; all that matters is that the web application just has to connect to port 4822 (where guacd listens by default) and use the Guacamole protocol. The architecture will take care of the rest.
Thankfully, the Java side of the Guacamole API provides simple classes which already implement the Guacamole protocol with the intent of tunneling it between guacd and the JavaScript half of your web application. A typical web application leveraging these classes needs only the following:
A class which extends
GuacamoleHTTPTunnelServlet, providing the tunnel between the JavaScript client (presumably using guacamole-common-js) and guacd.GuacamoleHTTPTunnelServletis an abstract class which is provided by the Guacamole API and already implements a fully functional, HTTP-based tunnel which the tunneling objects already part of guacamole-common-js are written to connect to. This class exists to make it easy for you to use Guacamole's existing and robust HTTP tunnel implementation.If you want to not use this class and instead use your own tunneling mechanism, perhaps WebSocket, this is fine; the JavaScript object mentioned above implements a common interface which you can also implement, and the Guacamole JavaScript client which is also part of guacamole-common-js will happily use your implementation as long as it provides that interface.
A web page which includes JavaScript files from guacamole-common-js and uses the client and tunnel objects to connect back to the web application.
The JavaScript API provided by the Guacamole project includes a full implementation of the Guacamole protocol as a client, implementations of HTTP and WebSocket-based tunnels, and mouse/keyboard/touch input abstraction. Again, as the Guacamole protocol and all parts of the architecture are documented here, you don't absolutely need to use these objects, but it will make your life easier. Mouse and keyboard support in JavaScript is finicky business, and the Guacamole client provided is well-known to work with other components in the API, being the official client of the project.
That's really all there is to it.
If you want authentication, the place to implement that would be in your extended
version of GuacamoleHTTPTunnelServlet; this is what the Guacamole
web application does. Besides authentication, there are many other things you could wrap
around your remote desktop application, but ultimately the base of all this is simple:
you have a tunnel which allows the JavaScript client to communicate with guacd, and you
have the JavaScript client itself, with the hard part already provided within
guacamole-common-js.
As with most tutorials, this tutorial begins with creating a project skeleton that establishes a minimal base for the tutorial to enhance in subsequent steps.
This tutorial will use Maven, which is the same build system used by the upstream Guacamole project. As the Guacamole project has a Maven repository for both the Java and JavaScript APIs, writing a Guacamole-based application using Maven is much easier; Maven will download and use the Guacamole API automatically.
All Maven projects must have a project descriptor, the
pom.xml file, in the root directory of the project. This
file describes project dependencies and specific build requirements. Unlike other
build tools like Apache Ant or GNU Autotools, Maven chooses convention over
configuration: files within the project must be placed in specific locations, and
the project dependencies must be fully described in the pom.xml. If this is done,
the build will be handled automatically.
The basis of this Guacamole-driven web application will be a simple HTML file
which will ultimately become the client. While the finished product will have an
HTTP tunnel written in Java, we don't need this yet for our skeleton. We will create
a very basic, barebones Maven project containing only
index.html and a web application descriptor file,
web.xml. Once these files are in place, the project can be
packaged into a .war file which can be deployed to your servlet
container of choice (such as Apache Tomcat).
As this skeleton will contain no Java code, it has no dependencies, and no build
requirements beyond the metadata common to any Maven project. The
pom.xml is thus very simple for the time being:
<project xmlns="http://maven.apache.org/POM/4.0.0"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xsi:schemaLocation="http://maven.apache.org/POM/4.0.0
http://maven.apache.org/maven-v4_0_0.xsd">
<modelVersion>4.0.0</modelVersion>
<groupId>org.apache.guacamole</groupId>
<artifactId>guacamole-tutorial</artifactId>
<packaging>war</packaging>
<version>1.1.0</version>
<name>guacamole-tutorial</name>
<properties>
<project.build.sourceEncoding>UTF-8</project.build.sourceEncoding>
</properties>
</project>Before the project will build, there needs to be a web application deployment
descriptor, web.xml. This file is required by the Java EE
standard for building the .war file which will contain the web
application, and will be read by the servlet container when the application is
actually deployed. For Maven to find and use this file when building the
.war, it must be placed in the
src/main/webapp/WEB-INF/ directory.
<?xml version="1.0" encoding="UTF-8"?>
<web-app version="2.5"
xmlns="http://java.sun.com/xml/ns/javaee"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xsi:schemaLocation="http://java.sun.com/xml/ns/javaee
http://java.sun.com/xml/ns/javaee/web-app_2_5.xsd">
<!-- Basic config -->
<welcome-file-list>
<welcome-file>index.html</welcome-file>
</welcome-file-list>
</web-app>With the web.xml file in place and the skeleton
pom.xml written, the web application will now build
successfully. However, as the web.xml refers to a "welcome
file" called index.html (which will ultimately contain our
client), we need to put this in place so the servlet container will have something
to serve. This file, as well as any other future static files, belongs within
src/main/webapp.
For now, this file can contain anything, since the other parts of our Guacamole-driven web application are not written yet. It is a placeholder which we will replace later:
<!DOCTYPE HTML>
<html>
<head>
<title>Guacamole Tutorial</title>
</head>
<body>
<p>Hello World</p>
</body>
</html>Once all three of the above files are in place, the web application will build,
and can even be deployed to your servlet container. It won't do anything yet other
than serve the index.html file, but it's good to at least try
building the web application to make sure nothing is missing and all steps were
followed correctly before proceeding:
$mvn package[INFO] Scanning for projects... [INFO] ------------------------------------------------------------------------ [INFO] Building guacamole-tutorial [INFO] task-segment: [package] [INFO] ------------------------------------------------------------------------ ... [INFO] ------------------------------------------------------------------------ [INFO] BUILD SUCCESSFUL [INFO] ------------------------------------------------------------------------ [INFO] Total time: 4 seconds [INFO] Finished at: Fri Jan 11 13:04:11 PST 2013 [INFO] Final Memory: 18M/128M [INFO] ------------------------------------------------------------------------$
Assuming you see the "BUILD SUCCESSFUL" message
when you build the web application, there will be a new file,
target/guacamole-tutorial-1.1.0.war, which
can be deployed to your servlet container and tested. If you changed the name or
version of the project in the pom.xml file, the name of this new
.war file will be different, but it can still be found
within target/.
Once we have a functional web application built, the next step is to actually add the references to the Guacamole API and integrate a Guacamole client into the application.
Now that we're adding Guacamole components to our project, we need to modify
pom.xml to specify which components are being used, and
where they can be obtained. With this information in place, Maven will automatically
resolve dependencies and download them as necessary during the build.
Regarding the build process itself, there are two main changes: we are now going
to be using Java, and we need the JavaScript files from guacamole-common-js included
automatically inside the .war.
Guacamole requires at least Java 1.6, thus we must add a section to the
pom.xml which describes the source and target Java
versions:
...
<build>
<plugins>
<!-- Compile using Java 1.6 -->
<plugin>
<groupId>org.apache.maven.plugins</groupId>
<artifactId>maven-compiler-plugin</artifactId>
<version>3.3</version>
<configuration>
<source>1.6</source>
<target>1.6</target>
</configuration>
</plugin>
</plugins>
</build>
...Including the JavaScript files from an external project like guacamole-common-js requires using a feature of the maven war plugin called overlays. To add an overlay containing guacamole-common-js, we add a section describing the configuration of the Maven war plugin, listing guacamole-common-js as an overlay:
...
<build>
<plugins>
...
<!-- Overlay guacamole-common-js (zip) -->
<plugin>
<groupId>org.apache.maven.plugins</groupId>
<artifactId>maven-war-plugin</artifactId>
<version>2.6</version>
<configuration>
<overlays>
<overlay>
<groupId>org.apache.guacamole</groupId>
<artifactId>guacamole-common-js</artifactId>
<type>zip</type>
</overlay>
</overlays>
</configuration>
</plugin>
</plugins>
</build>
...With the build now configured, we still need to add dependencies and list the repositories those dependencies can be downloaded from.
As this is a web application which will use the Java Servlet API, we must explicitly include this as a dependency, as well as the Guacamole Java and JavaScript APIs:
...
<dependencies>
<!-- Servlet API -->
<dependency>
<groupId>javax.servlet</groupId>
<artifactId>servlet-api</artifactId>
<version>2.5</version>
<scope>provided</scope>
</dependency>
<!-- Main Guacamole library -->
<dependency>
<groupId>org.apache.guacamole</groupId>
<artifactId>guacamole-common</artifactId>
<version>1.1.0</version>
<scope>compile</scope>
</dependency>
<!-- Guacamole JavaScript library -->
<dependency>
<groupId>org.apache.guacamole</groupId>
<artifactId>guacamole-common-js</artifactId>
<version>1.1.0</version>
<type>zip</type>
<scope>runtime</scope>
</dependency>
</dependencies>
...The Java Servlet API will be provided by your servlet container, so Maven does not need to download it during the build, and it need not exist in any Maven repository.
With these changes, the web application will still build at this point, even though no Java code has been written yet. You may wish to verify that everything still works.
If the pom.xml was updated properly as described above, the
web application should build successfully, and the Guacamole JavaScript API should
be accessible in the guacamole-common-js/ subdirectory of your
web application after it is deployed. A quick check that you can access
/guacamole-tutorial-1.1.0/guacamole-common-js/all.min.js
is probably worth the effort.
As with the other tutorials in this book, we will keep this simple for the sake of demonstrating the principles behind a Guacamole-based web application, and to give developers a good idea of where to start looking when it's time to consult the API documentation.
It is the duty of the class extending
GuacamoleHTTPTunnelServlet to implement a function called
doConnect(). This is the only function required to be
implemented, and in general it is the only function you should implement; the other
functions involved are already optimized for tunneling the Guacamole
protocol.
The doConnect() function returns a
GuacamoleTunnel, which provides a persistent
communication channel for GuacamoleHTTPTunnelServlet to use
when talking with guacd and initiating a connection with some arbitrary remote
desktop using some arbitrary remote desktop protocol. In our simple tunnel, this
configuration will be hard-coded, and no authentication will be attempted. Any user
accessing this web application will be immediately given a functional remote
desktop, no questions asked.
Create a new file, TutorialGuacamoleTunnelServlet.java,
defining a basic implementation of a tunnel servlet class:
package org.apache.guacamole.net.example;
import javax.servlet.http.HttpServletRequest;
import org.apache.guacamole.GuacamoleException;
import org.apache.guacamole.net.GuacamoleSocket;
import org.apache.guacamole.net.GuacamoleTunnel;
import org.apache.guacamole.net.InetGuacamoleSocket;
import org.apache.guacamole.net.SimpleGuacamoleTunnel;
import org.apache.guacamole.protocol.ConfiguredGuacamoleSocket;
import org.apache.guacamole.protocol.GuacamoleConfiguration;
import org.apache.guacamole.servlet.GuacamoleHTTPTunnelServlet;
public class TutorialGuacamoleTunnelServlet
extends GuacamoleHTTPTunnelServlet {
@Override
protected GuacamoleTunnel doConnect(HttpServletRequest request)
throws GuacamoleException {
// Create our configuration
GuacamoleConfiguration config = new GuacamoleConfiguration();
config.setProtocol("vnc");
config.setParameter("hostname", "localhost");
config.setParameter("port", "5901");
config.setParameter("password", "potato");
// Connect to guacd - everything is hard-coded here.
GuacamoleSocket socket = new ConfiguredGuacamoleSocket(
new InetGuacamoleSocket("localhost", 4822),
config
);
// Return a new tunnel which uses the connected socket
return new SimpleGuacamoleTunnel(socket);;
}
}Place this file in the
src/main/java/org/apache/guacamole/net/example subdirectory
of the project. The initial part of this subdirectory,
src/main/java, is the path required by Maven, while the
rest is the directory required by Java based on the package associated with the
class.
Once the class defining our tunnel is created, it must be added to the
web.xml such that the servlet container knows which URL
maps to it. This URL will later be given to the JavaScript client to establish the
connection back to the Guacamole server:
...
<!-- Guacamole Tunnel Servlet -->
<servlet>
<description>Tunnel servlet.</description>
<servlet-name>Tunnel</servlet-name>
<servlet-class>
org.apache.guacamole.net.example.TutorialGuacamoleTunnelServlet
</servlet-class>
</servlet>
<servlet-mapping>
<servlet-name>Tunnel</servlet-name>
<url-pattern>/tunnel</url-pattern>
</servlet-mapping>
...The first section assigns a unique name, "Tunnel", to the servlet class we just
defined. The second section maps the servlet class by it's servlet name ("Tunnel")
to the URL we wish to use when making HTTP requests to the servlet:
/tunnel. This URL is relative to the context root of the web
application. In the case of this web application, the final absolute URL will be
/guacamole-tutorial-1.1.0/tunnel.
As the Guacamole JavaScript API already provides functional client and tunnel implementations, as well as mouse and keyboard input objects, the coding required for the "web" side of the web application is very minimal.
We must create a Guacamole.HTTPTunnel, connect it to our
previously-implemented tunnel servlet, and pass that tunnel to a new
Guacamole.Client. Once that is done, and the
connect() function of the client is called,
communication will immediately ensue, and your remote desktop will be
visible:
...
<body>
<!-- Guacamole -->
<script type="text/javascript"
src="guacamole-common-js/all.min.js"></script>
<!-- Display -->
<div id="display"></div>
<!-- Init -->
<script type="text/javascript"> /* <![CDATA[ */
// Get display div from document
var display = document.getElementById("display");
// Instantiate client, using an HTTP tunnel for communications.
var guac = new Guacamole.Client(
new Guacamole.HTTPTunnel("tunnel")
);
// Add client to display div
display.appendChild(guac.getDisplay().getElement());
// Error handler
guac.onerror = function(error) {
alert(error);
};
// Connect
guac.connect();
// Disconnect on close
window.onunload = function() {
guac.disconnect();
}
/* ]]> */ </script>
</body>
...If you build and deploy the web application now, it will work, but mouse and keyboard input will not. This is because input is not implemented by the client directly. The Guacamole.Client object only decodes the Guacamole protocol and handles the display, providing an element which you can add manually to the DOM. While it will also send keyboard and mouse events for you, you need to call the respective functions manually. The Guacamole API provides keyboard and mouse abstraction objects which make this easy.
We need only create a Guacamole.Mouse and
Guacamole.Keyboard, and add event handlers to handle
their corresponding input events, calling whichever function of the Guacamole client
is appropriate to send the input event through the tunnel to guacd:
...
<!-- Init -->
<script type="text/javascript"> /* <![CDATA[ */
...
// Mouse
var mouse = new Guacamole.Mouse(guac.getDisplay().getElement());
mouse.onmousedown =
mouse.onmouseup =
mouse.onmousemove = function(mouseState) {
guac.sendMouseState(mouseState);
};
// Keyboard
var keyboard = new Guacamole.Keyboard(document);
keyboard.onkeydown = function (keysym) {
guac.sendKeyEvent(1, keysym);
};
keyboard.onkeyup = function (keysym) {
guac.sendKeyEvent(0, keysym);
};
/* ]]> */ </script>
...At this point, we now have a fully functional Guacamole-based web application. This web application inherits all the core functionality present in the official Guacamole web application, including sound and video, without very much coding.
Extending this application to provide authentication, multiple connections per user, or a spiffy interface which is compatible with mobile is not too much of a stretch. This is exactly how the Guacamole web application is written. Integrating Guacamole into an existing application would be similar.