Research on Disaster Tolerance System Based on Network Dataflow

In recent years, the disaster tolerance capability of network information systems has become the focus of research interest. In this paper, a disaster tolerance system architecture based on network dataflow is proposed and several techniques have been studied, namely, the rapid dataflow capturing, the failure packet filtering, the packet sequential storage technique supporting rapid retrieval, and the dataflow replay key technique supporting any revisit point. The disaster tolerance system using data streaming method does not incur extra system overhead, and is independent of specific applications. It can improve the data disaster tolerance generality and reduce deployment costs.