Geo-Redundant DR Design

Geo-redundant DR can ensure service continuity.

• This solution ensures unparalleled service continuity and maximum data availability, safeguarding both data and critical services even during regional natural disasters.
• The RPO is determined by the database replication interval. Servers at the DR center are always running, so an RTO of almost zero can be achieved. The time needed to complete a DR switchover depends on how long the DNS cache takes to update. It usually takes a few minutes and can be faster if GSLB is used.
  Figure 1 Three-AZ HA design
  

Key design points:

• The production and DR centers are deployed in two different regions of Huawei Cloud.
• The production centers are deployed in two AZs, and the DR center is deployed in one AZ.
• RDS instances are deployed at the production and DR centers. One primary and two standby databases are deployed.
• Configurations, logs, snapshots, and backups generated at the production and DR centers are replicated across regions using OBS.
• If an AZ at the production site is faulty, services are switched to the other AZ, and switchover is performed between the primary/standby databases.
• If both production centers fail, a database switchover is performed, and by modifying DNS configuration, all user traffic is directed to the DR center.
• After the production centers recover, a database switchover is performed, and DNS directs all user traffic to the production centers.
• To improve DR center utilization, some read-only and analysis services are distributed to the DR center.

Establishing HA DR capabilities is a complex project that includes ingress traffic control, service layer reconstruction, middleware and database control, and the collaboration of numerous systems. It requires professional skills to build HA DR systems. If customers lack related experience and want to quickly build an HA DR system, the Multi-cloud high Availability Service (MAS) is a good choice. This service is derived from the multi-cloud application HA solution of Huawei consumer services. It provides end-to-end service failover and DR drill capabilities that include everything from the traffic ingress and application layer, to the data layer. MAS ensures quick service recovery and improved service continuity.