Marsh James, Glencross Mashhuda, Pettifer Steve, Hubbold Roger
School of Computer Science, The University of Manchester, UK.
IEEE Trans Vis Comput Graph. 2006 May-Jun;12(3):405-16. doi: 10.1109/TVCG.2006.40.
Network architectures for collaborative virtual reality have traditionally been dominated by client-server and peer-to-peer approaches, with peer-to-peer strategies typically being favored where minimizing latency is a priority, and client-server where consistency is key. With increasingly sophisticated behavior models and the demand for better support for haptics, we argue that neither approach provides sufficient support for these scenarios and, thus, a hybrid architecture is required. We discuss the relative performance of different distribution strategies in the face of real network conditions and illustrate the problems they face. Finally, we present an architecture that successfully meets many of these challenges and demonstrate its use in a distributed virtual prototyping application which supports simultaneous collaboration for assembly, maintenance, and training applications utilizing haptics.
协作虚拟现实的网络架构传统上一直由客户端-服务器和对等方法主导，在延迟最小化至关重要的情况下，通常青睐对等策略，而在一致性是关键的情况下，则采用客户端-服务器策略。随着行为模型日益复杂以及对更好的触觉支持的需求，我们认为这两种方法都无法为这些场景提供足够的支持，因此需要一种混合架构。我们讨论了在实际网络条件下不同分发策略的相对性能，并说明了它们所面临的问题。最后，我们提出了一种成功应对诸多此类挑战的架构，并展示了其在分布式虚拟原型应用中的应用，该应用支持利用触觉进行装配、维护和培训应用的同步协作。
