Many applications require a multicast communication service to distribute data or coordinate . This service can be delivered at various layers of the OSI stack, the most common ones being the network layer (IP-multicast) and as middleware in which case it is known as application layer multicasting. Depending on the application, emphasis is on different quality attributes ranging from performance and scalability, e.g. for content delivery networks, to reliability, e.g. in a group of replica managers. The following references concern IP-multicasting

• The evolution of multicast: from the MBone to interdomain multicast to Internet2 deployment
• An Architecture for Wide-Area Multicast Routing
• Deployment issues for the IP multicast service and architecture

and the next references concern application layer multicasting

• Scalable Application Layer Multicast
• Overcast: Reliable Multicasting with an Overlay Network
• Scribe: A Large-Scale and Decentralized Application-Level Multicast Infrastructure
• Reliable Multicast in Data Center Networks
• Large-scale cooperative caching and application-level multicast in multimedia content delivery networks

In your essay explain how each of the two approaches works. Address the architectural principles of their design, whether, and if so how, group management is provided and mention the respective pros and cons of their usage. It is sufficient to study one article out of each list in detail and use the others merely as background information. Furthermore, indicate which of the quality concerns mentioned above are realized and whether and what (experimental) evidence is given to support this. In particular, discuss metrics that are used to measure the quality of an overlay network.