A personal profiling approach for caching and prefetching in distributed virtual environment

Abstract

Over the last decade, continuous evolution on both Internet and multimedia technology has turned the aspiration to Distributed Virtual Environment (DVE) across the Internet into a reality. A virtual environment usually contains a huge number of virtual objects represented by object models which are stored in a database server. When a user navigates through the virtual environment, the required data is transmitted from the server to the client machine for display through the underlying network. In this thesis, we aim at improving the performance of a DVE system across the Internet, both in terms of responsiveness and visual quality, under the existing constraint of relatively low bandwidth of the Internet. We propose the multi-resolution caching mechanism for effective manipulation of client-side caching. Our proposed multi-resolution caching mechanism exploits the progressive characteristics of the recently proposed multi-resolution modeling to improve caching performance. The caching mechanism is further complemented by a user-profiling based prefetching mechanism to predict objects that will be accessed in the future. To quantify the performance of the proposed mechanisms, we perform a number of simulated experiments. The simulation results show that our proposed caching mechanism out-performs the traditional one and the prefetching mechanism is useful in reducing the response time of the system. To validate the simulation study, we implement a prototype. We conduct a series of experiments based on the prototype, along with appropriate quantitative analysis. From the results, we identify scenarios where our approach is beneficial, resulting in shorter response time, better cache hit ratio and reduced latency. These results demonstrate the effectiveness of our proposed mechanisms for use in DVE.