Network Adaptive Consistency Control Scheme Based On The Continuous Replication System To Achieve

With the development of the computer network, the distributed application architecture becomes more and more popular. Using replicated method to build distributed system can efficiently improve the responsiveness of the system and reduce network traffic. Hence, it is widely used in practice.Some replication-based distributed systems are discrete system. The entities of these systems can only change their states according to user operations. But more replication-based distributed systems are continuous systems. The entities of these systems change their states not only because of user operations, but also because of the passing of time. These are the so-called replicated continuous system. The replicated continuous system is more valuable and is widely used in computer games, distributed virtual environments and military simulations.Like other distributed systems, the replicated continuous system has the same consistency control issue. It is important to maintain consistency among sites of the replicated continuous system.Now there are two major approaches to maintain consistency of the replicated continuous system. They are DR algorithm and Local Lag. While these two approaches can solve the consistency issue of the replicated system, they do have some shortcomings. One is that they have not taken network status into good account. So when the status of network changes a lot. It is hard to take these approaches to maintain the consistency of the system. Based on that knowledge, an adaptive consistency control scheme on network is proposed, so that it is possible to maintain the consistency of the system even when the network is changing dynamically.This consistency control scheme that is used in the replicated continuous system and has the adaptive ability on network consists of three technologies, Local Lag, TimeWarp and network measurement. To maintain the system consistency, this scheme first uses network measurement techniques to monitor and measure the network. And the results are used to change the local lag. After doing this, Local Lag is used to control the consistency of the system. If short term inconsistency emerges, a TimeWarp mechanism will be used to repair the state of the system.To study and implement this consistency control scheme, the replicated continuous system is analyzed first in this paper. Then the two major consistency control technologies which are currently used in the replicated systems are introduced. And their drawbacks are pointed out. After this, the adaptive consistency control scheme on network is presented. This scheme is implemented and used in a networked game. To prove that the scheme does work, we test the game under some simulated network status and analyze the testing results. The final section of the paper does some summary and looking forward work.