Research and application of dual active disaster preparedness system based on redo read-write separation cluster

Abstract

At present, the security and reliability of information systems have directly related to the development of various businesses of the State Grid Corporation. In order to solve problems such as data security, business continuity, and performance pressure, and avoid the impact of system disruption or data loss caused by a disaster on the business, it is necessary to research the solution of active-active disaster recovery. In this paper, based on a thorough study of the principle of read-write separation and its technical advantages in solving the performance bottleneck of high-concurrency business systems, a read-write separation cluster based on redo is designed in the data layer of the active and standby systems, and at the same time, the information system The active layer, the application layer, the data layer, and the storage layer of the architecture are designed for active-active layer by layer and a set of remote active-active disaster recovery systems is constructed. Considering that Zhejiang Power’s OMS system is a highly concurrent business system and the standby system is a cold standby with full service capabilities, this design scheme is used to implement transformation and upgrade of Zhejiang Power’s OMS business system. Through comparative testing, the new system is capable of providing services after a disaster. Interrupted, data zero loss recovery capability, and system response speed increased by more than 50%.