Research Of Self-organization Multimedia Overlay-network

Overlay network is a virtual network which constructs a user level graph on top of an existing networking infrastructure such as the Internet, it is a user-level topology built by selecting a subset of the nodes in the underlying network and connecting them by virtual links. It can easily be employed and provide much better quality of service without requiring changes of standards and without the risks associated with experimentation on an operational network infrastructure. Spines is an open source overlay network system that allows easy deployment and testing of overlay protocols, which employs hop-by-hop reliability techniques considerably reduces the average latency and jitter of reliable communication. It supports real-time multi-media service which requires good QoS, and can be easily applied to development and test of network protocol, but less dynamic. P2P overlay network with dynamic self-organization functions, its dynamic better.This thesis focus on the incorporation of Spines and DHT technology, to build the new multi-media network SMON(Self-organization Multimedia Overlay Network) based on both the hop-by-hop network transmission and real-time to restore with Spines and dynamic self-organizing feature of DHT technology. This thesis first analyzes the Spines network software architecture, and its main advantages and disadvantages, DHT technology main elements and its advantages and disadvantages, on this basis this thesis analysis the feasibility of combining Spines network and the DHT technology, and then described SMON network on the design ideas and the specific design. By joining the DHT technology to enhance the dynamic, adding a copy of the dynamic redundancy management DRM (Dynamic Redundant Management), walking mechanism ERW(Enhanced Random Walks) to increase robustness and scalability, describes the network specific work process. Finally, used NS2 simulation tools to build SMON network simulation platform, and performance comparison about SMON network and the Spines network in the dynamic, resource efficiency and load balancing. By theoretical analysis and simulation results shows that the structure be able to adapt to the dynamic changes in network topology, with accession and left of the nodes for routing update, compared with the original Spines network have not spending a lot business, and has good self-organized performance.