Disaster Recovery in the Hybrid Cloud: Enabling High Availability

As organizations increasingly adopt cloud-based solutions, it’s no surprise that a growing number of organizations are using a hybrid cloud model, including disaster recovery as a service (DRaaS), for disaster recovery. In this model, organizations host mission-critical applications on premises while outsourcing disaster recovery operations to a cloud services provider. The advantage of this strategy is that organizations maintain control of data management even as they offload the capital expenses and operational burdens associated with a standby disaster recovery site.

Although the hybrid cloud for DR gives organizations important advantages, it still has one serious drawback: It cannot guarantee high availability (HA) when disaster strikes, because (without Axxana’s Phoenix) data loss will always occur. That’s because synchronous replication to the cloud is not feasible; the latency and high cost of communication lines associated with synchronous replication between the primary site and the distant DR site are too great.  In the event of a disaster in a typical hybrid DR model, operations use asynchronous replication to fail over to the standby site. In this scenario, the organization will lose data, causing availability and business continuity to plummet while the organization races to recover and reconcile transactions.

Enabling High Availability in Hybrid DR Environments

The good news is that Axxana’s Phoenix offers a solution for achieving high availability in hybrid cloud DR environments. The solution leverages Axxana’s Phoenix for Oracle Far Sync, standard Oracle components, and creation of a dedicated Oracle service to automate application failover.

To arrive at the solution, Axxana engineers conducted research to assess options for plugging Axxana’s Phoenix for Oracle Far Sync into Oracle RDBMS-based application stacks in order to automate failover and achieve extremely high availability without requiring human intervention or extensive changes to code—and without impacting the application running in the primary site.

The main components of the proposed solution, shown in Figure 1, include:

  • Phoenix for Oracle Far Sync
  • Oracle Broker and its Observer component
  • A dedicated Oracle service that manages failover under a single session; the service detects a disaster, performs the failover, and replays the failed transactions, and does so in a transparent manner that practically eliminates impact on the application

Figure 1- HA over DR in the Hybrid Cloud

Most applications require complex coding to ensure that high-availability solutions address every possible disaster scenario; however, using the solution, organizations do not have to develop complex code for each new application. Once the underlying service configuration is created, organizations can achieve both disaster recovery and high availability on any application out of the box. Essentially, the service layer is application agnostic. It manages high availability and application failover without requiring staff to lift a finger. Using the solution to master seamless application failover, organizations can boost availability even higher to meet aggressive service level agreements (SLAs), while alleviating staffing burdens related to both development and management.

For in-depth technical information about Axxana’s high availability solution for the hybrid cloud, read the white paper: Disaster Recovery and High Availability in the Hybrid Cloud.