A UML 2.0 Communication Diagram models the objects or parts of a system, the interactions (or messages) between them, and the sequence in which these interactions occur.  There are a lot of similarities between communication diagrams and sequence diagrams in terms of the information they show, but because of how each diagram presents the information, one diagram may be better at conveying or emphasizing specific information over the other.

Communication diagrams use a free-form arrangement of objects or parts of a system.  This can be compared to how classes and objects are laid out in UML class and object diagrams.  Then the interactions between the objects or parts of the system are show and labeled to indicate the chronological sequence in which they occur.  The free-form arrangement of objects lends itself wekk to showing the sequenced interactions in a more compact space, allowing the analyst to place objects that have the highest number of interactions with each other near one another.  This is the advantage of the communication diagram over the sequence diagram.  While showing nearly all of the same information as a sequence diagram, the communication diagram can, at a glance, place a strong emphasize on which objects are interacting with one another

While communication diagrams are formally intended to show system objects and the interactions between them, many analysts choose to create them at a higher level of abstraction.  Instead of showing the interactions between objects of a system, larger parts of a system may be represented such as the interaction between web methods, web services, or entire systems. 

By using the communication diagram in this way, it shows some similarities to a system context diagram.  The primary differences between the two are that a system context diagram places a focus on a single system in context along with which actors and systems outside of the scope of the system interact with it. Additionally, a system context diagram does not show the sequence of interactions.