Logo Computer scientist,
engineer, and educator
• Articles • Articles about computing • Articles about software development • Apache integration software

Creating an Apache Camel route in Java from the ground up

This article describes how to use the Apache Camel integration framework to carry out a simple message routing operation. The operation is specified in Java but, to be honest, there isn't much actual Java coding involved. In fact, the example I use is the same copy-files-between-directories sample found in many books and articles on Camel, with one difference — I describe a step-by-step procedure that uses nothing more than command-line tools and a text editor to accomplish the task. You don't need Maven, you don't need Ant, you don't need an IDE. Of course, I've got nothing against IDEs and sophisticated tools, but I do feel they are a hindrance to education. So much is done automatically that the developer is relieved of the need to understand anything deeply. Working with nothing more than a text editor and a command prompt, you really have to understand what's going on.

The specific command-line examples I give are for Linux, because that's what I use for development work. However, they should work on other platforms, provided you know how to run java and javac at the command line.

I'm assuming that the reader has a good working knowledge of Java, and at least a vague idea of how Apache Camel works.

Obtaining and setting up Camel

The latest release of Camel can be obtained from the Camel Web site. For this example I'm using the binary (not source) distribution for Linux, version 2.12.1.

To use Camel you'll also need the logging facade SLF4J (available here), and the logger log4j (available here). Again, binary distributions are fine for this example. In fact, you can manage without log4j, and just use the built-in logger in SLF4J; however this logged appears (so far as I can tell) to be non-configurable, which is not very helpful if you have complicated debugging needs.

Installation consists of nothing more than unpacking the Camel, SLF4J, and log4j binary bundles into convenient directores. There are no executables, just libraries, configuration, and samples. You need to note where the lib directories for these three components are, because you'll need them for building and running code.

It goes without saying that you'll need a Java compiler and JVM to run Camel; happily Camel is not hugely fussy about which Java version or vendor you use, provided it's reasonably up-to-date.

Interlude — Camel routes

The main purpose of Camel is to implement routes. A route is a flow of data from one or more endpoints, to one or more other endpoints, optionally with transformations and conversions in between. Camel has built-in support for FTP, JMS, and HTTP endpoints, among many others, as well as simple file readers and writers, which we'll be using in this example. Many different types of data and format conversin are also built in, but new ones can easily be added.

Camel allows routes to be specified in a variety of domain specific languages; it is possible to specify very sophisticated routing using just XML, for example. The advantage of using Java is that allows for customer data processing, as a later article will demonstrate.

The Camel distribution provides a runtime engine that processes routes according to the developer's specification. In this example, we will use this built-in engine although, very often, in production settings the Camel routing is integrated into a general-purpose OSGi container like Apache ServiceMix or Eclipse Equinox. A later article will demonstrate this also.

Coding and compiling the route

Here is the Java source code for our basic file copier route; it is a single Java source file called FileCopier.java.
Simple Camel file copier 

(c)2013 Kevin Boone
import org.apache.camel.*;
import org.apache.camel.impl.*;
import org.apache.camel.builder.*;

public class FileCopier
  public static void main(String args[]) throws Exception
    CamelContext context = new DefaultCamelContext();
    context.addRoutes(new RouteBuilder()
      public void configure()
        // Set up the route — from the intput directory, to
        //  the output directory, with no other processing
        from ("file:/tmp/in?noop=true").to("file:/tmp/out");
    // Start the Camel route

    // Wait five minutes, then stop
    Thread.sleep (60*5*1000);
    context.stop ();
This code obtains a CamelContext — a representation of the Camel routing engine — and adds a route to it. The route is specified in the form of an anonymous class that implements the RouteBuilder interface. The route itself consists of one line of code:
from ("file:/tmp/in?noop=true").to("file:/tmp/out");
which sets up an assication between two directory endpoints. The action happens from as soon as we call context.start() and continues until (you guessed it) the call to context.stop(). The relevant directories are /tmp/in and /tmp/out; of course you can change these if necessary.

To compile this route, we simply compile this Java source, with a class search path that references the camel-core JAR. For Camel version 2.12.1, the relevant JAR is called camel-core-2.12.1.jar, in the Camel lib directory.

To compile the class at the command line:

$ javac -classpath=/path/to/camel/lib/camel-core-2.12.1.jar FileCopier.java
This should produce FileCopier.class.

Testing the route

If you've successfully compiled the route class, it's time to test it. The route will copy files between a pair of directories; you can create these first if you like, but Camel will create them on demand if they do not exist.

Starting the route

If you want to be able to configure the amount of logging produced by Camel, I would recommend that you run the Camel route with log4j as the logging engine, and create a log4j.properties file to configure it. If you're not bothered about logging you can use the default logger supplied with SLF4J — just refer to the JAR file slf4j-simple.jar on the class search path, rather than the slf4j-log4j.jar in the examples below. If you do use log4j, the following log4j.properties file will enable debug-level logging for Camel, which you'll probably need at some point.
log4j.rootLogger=INFO, out
log4j.properties needs to be on the class search path somewhere, otherwise the logger will not be able to find it. Alternatively, you can add a system property to the java command line to indicate where this file is:
java -Dlog4j.configuration=/path/to/log4j.properties"
In general, to run the Camel route, the class search path will need to indicate: your own compiled classes, the log4j logger JAR, the SLF4J class JAR that interfaces to log4j, the SLF4J API JAR, and the core Camel JAR. For this simple example, the following command-line should do the trick, assuming that log4j.properties is in the current directory.
java \
  -classpath /path/to/log4j/log4j-1.2.17.jar:\
  /path/to/camel/lib/slf4j-api-1.6.6.jar:. FileCopier
Of course, this command should be on one long line; I've just split it up for clarity. You'll need to replace the /path/to... entries with the actual paths of this software components on your system, and possibly change the version numbers also.

Testing the route

To test the route, just copy a file to the 'in' directory. A short while later, it should appear in the 'out' directory. I've used a file called MYTEST; with debug-level logging, you should see output similar to the following in the console:
[-1) thread #0 - file:///tmp/in] FileConsumer                   
  DEBUG About to process file: GenericFile[/tmp/in/MYTEST] using exchange: Exchange[MYTEST]
[-1) thread #0 - file:///tmp/in] SendProcessor                  
  DEBUG Endpoint[file:///tmp/out] Exchange[MYTEST]
[-1) thread #0 - file:///tmp/in] FileOperations                 
  DEBUG Using FileChannel to write file: /tmp/out/MYTEST
[-1) thread #0 - file:///tmp/in] GenericFileProducer            
  DEBUG Wrote [/tmp/out/MYTEST] to [Endpoint[file:///tmp/out]]
[-1) thread #0 - file:///tmp/in] GenericFileOnCompletion        
  DEBUG Done processing file: GenericFile[/tmp/in/MYTEST] using exchange: Exchange[MYTEST]


There is, of course, a huge amount more to Camel than this. In practice, it's quite difficult to manage a complex Camel project (or any software project) using just the command line; hopefully, however, illustrating the individual steps like this has provided some better idea of what's actually happening at a technical level.


Source code bundle

Copyright © 1994-2013 Kevin Boone. Updated May 01 2014