With the large-scale natural disasters occurred in China in recent years, the disaster recovery mode of remote disaster recovery center city center with double "three center" also appears, this scheme has the ability of high availability and disaster recovery.

Double center city refers to the establishment of two can undertake key independent system operation data center in the city or near the city center, with double business processing ability is basically through the high-speed link data real-time synchronization, everyday situations can also share the business operation and management system, and switch operation in disaster situations; no loss of disaster emergency switch data, maintain business continuity. Compared with the remote disaster recovery mode, double city center has low investment cost, quick construction, operation and maintenance management is relatively simple, higher reliability advantages.

Remote disaster recovery center refers to the disaster recovery center in different places of the city to build a backup, backup for data center, when the reasons of the double center appeared such as natural disasters and the occurrence of fault, remote disaster recovery center can restore business with the backup data.

For the construction of Disaster Prevention Center three demand, HUAWEI Symantec Corp using Symantec software combined with the advantages of its own storage development capabilities, design the construction scheme of the typical scheme, principle and network as shown in fig..

 

Figure, city double switching application center, using Symantec Veritas VCS (Veritas Cluster Server) cluster software, production centers and disaster recovery center on the host host need to install VCS. VCS is used to monitor local, dual, or cluster status and to monitor status between local and remote clusters via component GCO (Global, Cluster, Option).

In the network layer, the city between the two center connected by optical fiber, high bandwidth between the two center, in response to the demand of real-time business data, the network between city offsite or IP wide area network can be realized, in order to save costs. The city center by fiber wavelength division multiplexing (WDM) technology for the construction of the two, only 1 or 2 fiber connection scenarioses, using WDM mode, to a virtual multiple FC or GE contact, meet the needs of the business and the data of multiple links between the two places. WDM technology can make full use of the huge bandwidth of optical fiber, greatly improve the system transmission capacity and reduce transmission costs. Therefore, it has been widely used in long-distance and backbone transmission of high-capacity transmission. With the introduction of WDM technology into metropolitan area network and access network, the whole network will become a seamless connection and provide support and connection for all kinds of different businesses. Therefore, WDM has great advantages in metropolitan area networks.

In the data storage layer, the deployment of Oceanspace VIS6000 or Oceanspace S5000/S6800E virtual storage disk array through the synchronous remote memory copy function synchronizes data copied to the disaster recovery site. Ensure production centers and disaster recovery center data completely consistent. Asynchronous replication function using the storage of the data through the wide area network is copied to the remote disaster recovery site, and ensure the integrity and availability of data. The role of the remote site is mainly used to prevent geographical and natural disasters. When the city's double center fails, it can ensure that there is a complete copy of the data in the other place, which is used for the recovery of the subsequent business.

The following describes the program specific data backup and business processes:

· Data backup 

The city center data using synchronous replication, the establishment of a copy of the data in the online update of city disaster recovery center. When the data is sent to the production center of the array, the array replication will also copy the data to a local disaster recovery center.

Asynchronous replication mode is adopted between the city disaster recovery center and remote disaster recovery center, regular data replication, asynchronous replication support incremental replication, data backup can save bandwidth, shorten the data backup time.

The VCS-Global Cluster on the host determines the availability of the resource by monitoring the status of the resource group, including database resource groups, network resource groups, and so on. The resource group state of online/offline/fault three, under normal circumstances, the production center when the resource group work state is online, and the resource group is offline state disaster recovery center. Each resource group can specify a running program or script at online and offline. When the program or script is executed, the resource group completes the online or offline process. When the detected production center fault state resources group, in the same city center production disaster recovery center will switch to ensure business continuity.

· Disaster tolerance switching 

The application of disaster tolerance switching consists of a series of action based on node components: stop disaster service, cut off the data replication link, establish data disaster recovery, disaster recovery start node baseline components, service notification front-end equipment business network switching. Specific actions can be combined with the actual situation, customized by script.

· Restore back to cut 

Cut back the work flow and the switching process of the principle is the same, just because switching is uncertain when triggered, may lead to part of business and influence; cut back through the manual confirmation, operation minimum under the condition (such as traffic is very small, or even suspend the business case). So cut back the recommended manual switching mode.

Application level disaster recovery uses automatic switching or manual switching, and users can modify the host cluster software during the deployment by configuring the switching configuration.

· City wide effective guarantee of data security and business continuity;

· Remote replication of data, according to the disaster situation, reduces the probability of data loss as much as possible;

· City dual Center for synchronous replication, real-time data synchronization, RPO=0;

· There is no distance restriction to ensure data consistency and ensure the effective protection of data;

· Remote disaster tolerance requires low bandwidth and advanced replication mechanism to improve bandwidth utilization.