Today, the telecommunication industry is undergoing two important developments with implications on future architectural solutions. These are the irreversible move towards Internet Protocol (IP)-based networking and the deployment of broadband access. Taken together, these developments offer the opportunity for more advanced and more bandwidth-demanding multimedia applications and services, e. g., IP television (IPTV), Voice over IP (VoIP) and online gaming. A plethora of Quality of Service (QoS) requirements and facilities are associated with these applications, e. g., multicast facilities, high bandwidth and low delay/jitter. Moreover, the architectural solution must be a unified one, and be independent of the access network and content management. An interesting solution to these challenges is given by overlay multicast networks. The goal of these networks is to create and to maintain efficient multicast topologies among the multicast participants as well as to minimize the performance penalty involved with application layer multicasting. Since they operate at the application layer, they suffer from two main drawbacks: higher delay and less efficient bandwidth utilization. It is therefore important to assess the performance of overlay multicast networks in “real- world”-like conditions. For this purpose, we first performed an in-depth measurement and modeling study of the packet delay at the network layer. The reported results are in the form of several important statistics regarding processing and queueing delays of a router. New results have been obtained that indicate that the delay in IP routers shows heavy-tailed characteristics, which can be well modeled with the help of several distributions, in the form of a single distribution or as a mixture of distributions. There are several components contributing to the delay in routers, i. e., processing delay, queueing delay and service time. It was observed that the component delay distribution that is most heavy-tailed has a decisive influence on delay. Furthermore, we selected three representative categories of overlay multicast networks for study, namely Application Level Multicast Infrastructure (ALMI), Narada and NICE is the Internet Cooperative Environment (NICE). The performance of these overlay multicast protocols was evaluated through a comprehensive simulation study with reference to a detailed set of performance metrics that captured application and network level performance. A particular interest was given to the issues of scalability, protocol dynamics and delay optimization as part of a larger problem of performance-aware optimization of the overlay networks. The simulations were configured to emulate “real-world”-like characteristics by implementing a heavy-tailed delay at the network level and churn behavior of the overlay nodes. A detailed analysis of every protocol is provided with regard to their performance. Based on our study, significant conclusions can be drawn regarding the scalability of the protocols with reference to overlay multicast group management, resource usage and robustness to churn. These results contribute to a deeper understanding of the requirements for such protocols targeted at, e. g., media streaming.