Multi-node and 2-storage solution for regional health information platform

Project introduction

To implement national and provincial health policies, laws and regulations, and to promote the medical and health system reform, the Municipal Health Bureau of one city plans to build a regional health information platform. The regional health information platform aims to build a municipal hospital management system data center. It revolves around construction of residents' health records information system, and sharing of hospital information system, disease prevention and control information system, health supervision information system data information, to build a data hosted network. Upward, it connects the Hainan Provincial Health Department. Downward, it extends to municipal medical and health institutions of various levels. Horizontally, it links to social security, family planning and other relevant departments. This could relate and store residents' health information, public health work information and health diagnosis information, as well as their data sharing and information utilization .

Regional health information platform collects a large quantity of health data, which is convenient for municipal health bureau to process and analyze huge amounts of data, with data center as the core, to find colonial high incidence diseases timely, and then give corresponding guide and start the control mechanism for epidemiology targetedly, by making use of business intelligence technologies including online analytical processing, data mining and data exhibition.

In health information construction of municipal health bureau, the regional health information platform plays a particularly important role. While building the platform, the disaster recovery solution that maximally ensures data security and application continuity should be planned.

 Disaster recovery demand

Oracle database is the background database of municipal health bureau's regional health information platform, and it is of top priority in entire health information platform. If the regional health information platform encounters data loss or business unavailability, it will severely impact the normal medical work of the whole city, affect residents' timely diagnosis and treatment, and also may cause great social influence.

Municipal health bureau information dept proposes to build high-level disaster solution for Oracle database, with requirement that the disaster recovery effect with RPO=0 and RTO≈0 should be realized. Whether it is a hardware failure (disk array, server), or software failures (database, operating system), it would not affect the normal running of business system, and database has zero data loss, and the entire solution achieves completely redundant protection.

For maintenance and management, the municipal health bureau puts forward that it should ensure health information platform to run safely and faultlessly, at the same time, disaster recovery solution must be convenient for system administrator’s routine management and maintenance.

The solution

After detailed demand analysis and research, the municipal health bureau adopted Rose Multi-node and 2-storage solution to build completely redundant architecture and realize RPO=0 and RTO≈0 effect.

Multi-node and 2-storage solution is completely redundant architecture, thoroughly solves the software and hardware single point of failure. Any single point of failure of servers, storage switches or disk array would not affect the normal running of the Oracle; it ensures zero data loss of Oracle database, and database service won't be interrupted.

At server level, it builds a redundant active-standby cluster. Rose solution monitors running state of servers and application resources in real time. If it encounters a failure, it will automatically switch the resources to standby server.

At storage level, it builds the redundant FC SAN storage architecture. On server side, it restores redundant storage path via Rose solution, and to build the LUN of 2 storages as RAID1. The dual active storage architecture has realized the synchronous data writes, one IO would be written synchronously into two storages. If any storage fails, database data would not be lost, so the database service would not be affected.

And, after it recovers from single storage failure, in the process to restore RAID1, Rose adopts difference synchronization technology to synchronize only data differences after single storage failure, which improves the efficiency of data synchronization, and is applicable for big data environment planned by municipal health bureau. Rose automatically recovers RAID1, and auto failover in case of application fault, with no need of human supervision, which greatly simplifies the maintenance and management work for RAID1 and servers, and reduces the management cost.

 Solution value

Rose multi-node and 2-storage solution meets the disaster recovery demand of the municipal health bureau, makes sure the regional health information platform to run normally, and has realized data-level and application-level disaster recovery.

After actual use for over 1 year,  Rose multi-node and 2-storage solution protects the regional health information platform effectively, and realize RPO=0 and RTO≈0, to provide the technical disaster recovery support for municipal health information construction .