A quantitative analysis of Local Data Management for 3D content streaming

Abstract

Content streaming is a mechanism used to distribute static geometry data to users of a virtual environment in real-time. Although most existing content streaming mechanisms have shown to meet their run-time performance requirements, they have an underlying constraint on the quality of 3D data that can be streamed. This restriction stems from the fact that most existing content streaming mechanisms require the server to transmit the same content every time the user encounters it. As a result, unnecessary bandwidth consumption would be induced. In this paper, we present a caching approach termed as the Local Data Management (LDM) framework to reduce bandwidth usage by maintaining local copies of geometry data in the client system. We also evaluate the performance of the LDM approach in conjunction with multiple multiresolution interest matching algorithms. Our goal is to compare the cost of strategies studied under various system conditions and scenarios.