The proliferation of the Internet has driven the development of various multicast applications such as network games, networked virtual environments, video/audio conferencing, collaboration and distributed interactive simulation. The limited deployment of IP multicast and the increased interest of the multi-party applications promotes the use of application layer multicast.
In application layer multicast, multicast routing is implemented and operated at an end-system as an application. However, end systems participating in an application layer multicast session are not designed to be optimized for packet delivery and heterogeneous in terms of computing capacity and network bandwidth. The delay incurred at end systems can contribute to the increase of the end-to-end delay.
In this thesis, we propose a model that explains the delay incurred at an end system in an application layer multicast session. And also, we propose a session throughput enhancement scheme based on a mesh among members incorporating the heterogeneous delay incurred at end systems. We analyze the impact of the heterogeneous delay incurred at end systems on the performance of the existing schemes with the proposed model and show the throughput enhancement of the proposed scheme.