This figure shows a geographically distributed disaster recovery configuration, which consists of one production site in San Francisco and one standby site in Atlanta. Data and configuration at the production site is synchronized to the standby site. If the production site fails, the standby takes over the handling of requests.