Stable and Scalable Method Based on Application-Level for Improving Distributed Interactive Visual Applications' Performance

Abstract

virtual reality application such as second life and other multi-player game is a set of special Distributed Interactive Applications (DIAs), because a sender node forwards data to receivers due to their respective priorities. In general application level protocol is adopted to multicast: the significance-based directed minimum spanning tree was designed for these DIAs. In this paper we propose a novel application level algorithm: Quantificational Analysis and Prediction for Significance with directed minimum Spanning Tree (QAPSST), which can efficiently predict priorities for the receivers and quantize the predicted priorities to build a multicast distribution tree data structure. Furthermore, QAPSST can easily integrate the quantized significance into game environment and simplify significance deployment. Our significance-based directed minimum spanning tree has significance-efficient predict mechanism and the system consumes a tremendous amount of resource and still is stable and scalable when its size increases drastically. The experiment results show that QAPSST is able to efficiently make significance predict and keep system stable with huge amount of users.