Research And Implementation Of The High-Availability Remote Disaster Recovery System

"9.11" incident makes everybody pay more attention on their application systems. Traditional data backup technologies and server cluster technologies have avoided some destructions in a certain level, which are caused by a variety of hardware and software failures, operational errors and viruses, and these technologies have protected data and system security. However, when facing large-scale disasters and emergencies, such as earthquakes, fires, terrorist attacks, war and so on, these technologies are powerless. Therefore, remote disaster recovery system is in necessary to rapidly recovery system data and keep the information systems playing normally.Based on the application background of the military command information database system, this paper firstly analyzes the current remote disaster recovery technology, and then provides an abstract model of the disaster recovery system. On the base of these works, the architecture of the remote disaster recovery system with high availability is designed. To improve the high availability, we focus on many technologies, such as remote load balancing, self-adaptive loading technology, data consistency, high reliable data coding algorithms and IP-SAN security authentication mechanism. Finally we realize a prototype system of remote disaster recovery, and provide some tests to the corresponding functional models. The concrete works are as follows.1. Analyze and compare the current main technologies on disaster recovery, and point out their deficiencies if they were applied in the military command information database system.2. To satisfy the special requirements of the military command information database system, an abstract model of the remote disaster recovery system is provided, further more, the architecture of the high-availability remote disaster recovery system is designed. At the same time, the functional modules are analyzed in detail especially.3. In order to ensure the high-availability of the remote disaster recovery system, we spare no effort to research load balancing technology to improve system performance; and to research self-adaptive loading technology to deal with the loading bottleneck when processing massive data; and to research data consistency technology to maintain the consistency between local data and remote data; and to research high-reliable data coding algorithm to improve the reliability of data transmission and finally research the IP-SAN certification mechanism to improve data transmission security.4. In order to test the feasibility of these key technologies, we realize the prototype system of remote disaster recovery, and test some key function modules. And the results show that the remote disaster recovery theory is feasible when taking into practice.