TY - GEN
T1 - Towards multi-site collaboration in 3D tele-immersive environments
AU - Wu, Wanmin
AU - Yang, Zhenyu
AU - Gupta, Indranil
AU - Nahrstedt, Klara
PY - 2008
Y1 - 2008
N2 - 3D tele-immersion (3DTI) has recently emerged as a new way of video-mediated collaboration across the Internet. Unlike conventional 2D video-conferencing systems, it can immerse remote users into a shared 3D virtual space so that they can interact or collaborate "virtually". However, most existing 3DTI systems can support only two sites of collaboration, due to the huge demand of networking resources and the lack of a simple yet efficient data dissemination model. In this paper, we propose to use a general publish-sub scribe model for multi-site 3DTI systems, which efficiently utilizes limited network resources by leveraging user interest. We focus on the overlay construction problem in the publish-sub scribe model by exploring a spectrum of heuristic algorithms for data dissemination. With extensive simulation, we identify the advantages of a simple randomized algorithm. We present optimization to further improve the randomized algorithm by exploiting semantic correlation. Experimental results demonstrate that we can achieve an improvement by a factor of five.
AB - 3D tele-immersion (3DTI) has recently emerged as a new way of video-mediated collaboration across the Internet. Unlike conventional 2D video-conferencing systems, it can immerse remote users into a shared 3D virtual space so that they can interact or collaborate "virtually". However, most existing 3DTI systems can support only two sites of collaboration, due to the huge demand of networking resources and the lack of a simple yet efficient data dissemination model. In this paper, we propose to use a general publish-sub scribe model for multi-site 3DTI systems, which efficiently utilizes limited network resources by leveraging user interest. We focus on the overlay construction problem in the publish-sub scribe model by exploring a spectrum of heuristic algorithms for data dissemination. With extensive simulation, we identify the advantages of a simple randomized algorithm. We present optimization to further improve the randomized algorithm by exploiting semantic correlation. Experimental results demonstrate that we can achieve an improvement by a factor of five.
UR - http://www.scopus.com/inward/record.url?scp=51849154967&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=51849154967&partnerID=8YFLogxK
U2 - 10.1109/ICDCS.2008.47
DO - 10.1109/ICDCS.2008.47
M3 - Conference contribution
AN - SCOPUS:51849154967
SN - 9780769531724
T3 - Proceedings - The 28th International Conference on Distributed Computing Systems, ICDCS 2008
SP - 647
EP - 654
BT - Proceedings - The 28th International Conference on Distributed Computing Systems, ICDCS 2008
T2 - 28th International Conference on Distributed Computing Systems, ICDCS 2008
Y2 - 17 July 2008 through 20 July 2008
ER -