An Adaptive Control Mechanism for Management of Networked

Abstract

To automatically maintain appropriate level of interested information for a large number of participants in networked virtual environments is of paramount importance. In this paper we propose an adaptive control mechanism to manage the information consistency and control the dynamic shared state for a large-scale virtual environment. The approach uses three major techniques for dynamic shared state maintenance - the shared repository technique, blind broadcasting method, and dead reckoning model to throttle the data transmission over the network. Depending upon the system load, the approach can automatically switch from one level to another level to maintain maximum performance in server. Performance evaluation on the proposed adaptive control mechanism, the tree-level consistency approach, and the shared repository technique were conducted. Performance evaluation results show that both the proposed mechanism and the tree-level consistency approach can maintain low system load over a wide range of participants, but the adaptive control mechanism can adapt to the state of server load and provide a better service quality.