There are many, many technologies in the world that have "Inter-process communication" or "networking" in their stated purpose: CORBA, DCE, DCOM, DCOP, XML-RPC, SOAP, MBUS, Internet Communications Engine (ICE), and probably hundreds more. Each of these is tailored for particular kinds of application. D-Bus is designed for two specific cases:

For the within-desktop-session use case, the GNOME and KDE desktops have significant previous experience with different IPC solutions such as CORBA and DCOP. D-Bus is built on that experience and carefully tailored to meet the needs of these desktop projects in particular. D-Bus may or may not be appropriate for other applications; the FAQ has some comparisons to other IPC systems.

The problem solved by the systemwide or communication-with-the-OS case is explained well by the following text from the Linux Hotplug project:

D-Bus may happen to be useful for purposes other than the one it was designed for. Its general properties that distinguish it from other forms of IPC are:

Your programming framework probably defines what an "object" is like; usually with a base class. For example: java.lang.Object, GObject, QObject, python's base Object, or whatever. Let's call this a native object.

The low-level D-Bus protocol, and corresponding libdbus API, does not care about native objects. However, it provides a concept called an object path. The idea of an object path is that higher-level bindings can name native object instances, and allow remote applications to refer to them.

The object path looks like a filesystem path, for example an object could be named /org/kde/kspread/sheets/3/cells/4/5. Human-readable paths are nice, but you are free to create an object named /com/mycompany/c5yo817y0c1y1c5b if it makes sense for your application.

Namespacing object paths is smart, by starting them with the components of a domain name you own (e.g. /org/kde). This keeps different code modules in the same process from stepping on one another's toes.

Each object has members; the two kinds of member are methods and signals. Methods are operations that can be invoked on an object, with optional input (aka arguments or "in parameters") and output (aka return values or "out parameters"). Signals are broadcasts from the object to any interested observers of the object; signals may contain a data payload.

Both methods and signals are referred to by name, such as "Frobate" or "OnClicked".

Each object supports one or more interfaces. Think of an interface as a named group of methods and signals, just as it is in GLib or Qt or Java. Interfaces define the type of an object instance.

DBus identifies interfaces with a simple namespaced string, something like org.freedesktop.Introspectable. Most bindings will map these interface names directly to the appropriate programming language construct, for example to Java interfaces or C++ pure virtual classes.

A proxy object is a convenient native object created to represent a remote object in another process. The low-level DBus API involves manually creating a method call message, sending it, then manually receiving and processing the method reply message. Higher-level bindings provide proxies as an alternative. Proxies look like a normal native object; but when you invoke a method on the proxy object, the binding converts it into a DBus method call message, waits for the reply message, unpacks the return value, and returns it from the native method..

In pseudocode, programming without proxies might look like this:

          Message message = new Message("/remote/object/path", "MethodName", arg1, arg2);
          Connection connection = getBusConnection();
          connection.send(message);
          Message reply = connection.waitForReply(message);
          if (reply.isError()) {
             
          } else {
             Object returnValue = reply.getReturnValue();
          }
        

Programming with proxies might look like this:

          Proxy proxy = new Proxy(getBusConnection(), "/remote/object/path");
          Object returnValue = proxy.MethodName(arg1, arg2);
        

When each application connects to the bus daemon, the daemon immediately assigns it a name, called the unique connection name. A unique name begins with a ':' (colon) character. These names are never reused during the lifetime of the bus daemon - that is, you know a given name will always refer to the same application. An example of a unique name might be :34-907. The numbers after the colon have no meaning other than their uniqueness.

When a name is mapped to a particular application's connection, that application is said to own that name.

Applications may ask to own additional well-known names. For example, you could write a specification to define a name called com.mycompany.TextEditor. Your definition could specify that to own this name, an application should have an object at the path /com/mycompany/TextFileManager supporting the interface org.freedesktop.FileHandler.

Applications could then send messages to this bus name, object, and interface to execute method calls.

You could think of the unique names as IP addresses, and the well-known names as domain names. So com.mycompany.TextEditor might map to something like :34-907 just as mycompany.com maps to something like 192.168.0.5.

Names have a second important use, other than routing messages. They are used to track lifecycle. When an application exits (or crashes), its connection to the message bus will be closed by the operating system kernel. The message bus then sends out notification messages telling remaining applications that the application's names have lost their owner. By tracking these notifications, your application can reliably monitor the lifetime of other applications.

Bus names can also be used to coordinate single-instance applications. If you want to be sure only one com.mycompany.TextEditor application is running for example, have the text editor application exit if the bus name already has an owner.

Applications using D-Bus are either servers or clients. A server listens for incoming connections; a client connects to a server. Once the connection is established, it is a symmetric flow of messages; the client-server distinction only matters when setting up the connection.

If you're using the bus daemon, as you probably are, your application will be a client of the bus daemon. That is, the bus daemon listens for connections and your application initiates a connection to the bus daemon.

A D-Bus address specifies where a server will listen, and where a client will connect. For example, the address unix:path=/tmp/abcdef specifies that the server will listen on a UNIX domain socket at the path /tmp/abcdef and the client will connect to that socket. An address can also specify TCP/IP sockets, or any other transport defined in future iterations of the D-Bus specification.

When using D-Bus with a message bus daemon, libdbus automatically discovers the address of the per-session bus daemon by reading an environment variable. It discovers the systemwide bus daemon by checking a well-known UNIX domain socket path (though you can override this address with an environment variable).

If you're using D-Bus without a bus daemon, it's up to you to define which application will be the server and which will be the client, and specify a mechanism for them to agree on the server's address. This is an unusual case.

Pulling all these concepts together, to specify a particular method call on a particular object instance, a number of nested components have to be named:

          Address -> [Bus Name] -> Path -> Interface -> Method
        

The bus name is in brackets to indicate that it's optional -- you only provide a name to route the method call to the right application when using the bus daemon. If you have a direct connection to another application, bus names aren't used; there's no bus daemon.

The interface is also optional, primarily for historical reasons; DCOP does not require specifying the interface, instead simply forbidding duplicate method names on the same object instance. D-Bus will thus let you omit the interface, but if your method name is ambiguous it is undefined which method will be invoked.

D-Bus works by sending messages between processes. If you're using a sufficiently high-level binding, you may never work with messages directly.

There are 4 message types:

A method call maps very simply to messages: you send a method call message, and receive either a method return message or an error message in reply.

Each message has a header, including fields, and a body, including arguments. You can think of the header as the routing information for the message, and the body as the payload. Header fields might include the sender bus name, destination bus name, method or signal name, and so forth. One of the header fields is a type signature describing the values found in the body. For example, the letter "i" means "32-bit integer" so the signature "ii" means the payload has two 32-bit integers.

A method call in DBus consists of two messages; a method call message sent from process A to process B, and a matching method reply message sent from process B to process A. Both the call and the reply messages are routed through the bus daemon. The caller includes a different serial number in each call message, and the reply message includes this number to allow the caller to match replies to calls.

The call message will contain any arguments to the method. The reply message may indicate an error, or may contain data returned by the method.

A method invocation in DBus happens as follows:

The bus daemon never reorders messages. That is, if you send two method call messages to the same recipient, they will be received in the order they were sent. The recipient is not required to reply to the calls in order, however; for example, it may process each method call in a separate thread, and return reply messages in an undefined order depending on when the threads complete. Method calls have a unique serial number used by the method caller to match reply messages to call messages.

A signal in DBus consists of a single message, sent by one process to any number of other processes. That is, a signal is a unidirectional broadcast. The signal may contain arguments (a data payload), but because it is a broadcast, it never has a "return value." Contrast this with a method call (see the section called “Calling a Method - Behind the Scenes”) where the method call message has a matching method reply message.

The emitter (aka sender) of a signal has no knowledge of the signal recipients. Recipients register with the bus daemon to receive signals based on "match rules" - these rules would typically include the sender and the signal name. The bus daemon sends each signal only to recipients who have expressed interest in that signal.

A signal in DBus happens as follows:

D-Bus objects may support the interface org.freedesktop.DBus.Introspectable. This interface has one method Introspect which takes no arguments and returns an XML string. The XML string describes the interfaces, methods, and signals of the object. See the D-Bus specification for more details on this introspection format.