A modelling and networking architecture for distributed virtual environments with multiple servers
Virtual Environments (VEs) attempt to give people the illusion of immersion that they are in a computer generated world. VEs allow people to actively participate in a synthetic environment. They range from a single-person running on a single computer, to multiple-people running on several computers connected through a network. When VEs are distributed on multiple computers across a network, we call this a Distributed Virtual Environment (DVE). Virtual Environments can benefit greatly from distributed strategies. A networked VE system based on the Client-Server model is the most commonly used paradigm in constructing DVE systems. In a Client-Server model, data can be distributed on several server computers. The server computers provide services to their own clients via networks. In some client-server models, however, a powerful server is required, or it will become a bottleneck. To reduce the amount of data and traffic maintained by a single server, the servers themselves can be distributed, and the virtual environment can be divided over a network of servers. The system described in this thesis, therefore, is based on the client-server model with multiple servers. This grouping is called a Distributed Virtual Environment System with Multiple- Servers (DVM). A DVM system shows a new paradigm of distributed virtual environments based on shared 3D synthetic environments. A variety of network elements are required to support large scale DVM systems. The network is currently the most constrained resource of the DVM system. Development of networking architectures is the key to solving the DVM challenge. Therefore, a networking architecture for implementing a DVM model is proposed. Finally, a DVM prototype system is described to demonstrate the validity of the modelling and network architecture of a DVM model.