A framework to practically constitute a Distributed Virtual Reality (DVR) system on heterogeneous network environment is shown. The proposed approach essentially integrate "haptic" channel in addition to the traditional interaction modalities such as vision and acoustic senses. The haptic interface technology adds the sense-of-touch and empowers the reality of DVR system by allowing users to touch the virtual objects. The human touch sensation, however, is very sensitive for delays and jitters in the networked virtual environment. Therefore, we propose a new DVR architecture to realize a practical system on a long-haul international network. We take a hybrid approach to realize both data consistency by the client-server and scalability by the peer-to-peer models. Some preliminary experiments using the Korea-Japan high-speed research networks to validate the proposed method are also mentioned.