Forrester data indicates that improving business continuity and disaster recovery (BC/DR) ranks at or near the top of Asia Pacific companies’ vendor and product selection criteria. However, the technology infrastructure to support BC/DR is poorly architected, funded, and managed. Many firms are now looking to potentially outsource their DR infrastructure and leverage the public cloud’s pay-per-use model via vendors like Amazon Web Services.

Despite the low price points, high degree of automation, and robust infrastructure, few companies are using public cloud as a DR site. This will change as network carrier infrastructure matures and firms start complementing existing infrastructure with public cloud initiatives. Firms will discover that cloud DR, particularly in the public cloud, is an opportunity to free up resources and capital to support new business initiatives. Forrester sees three primary use cases for cloud DR:

  • A milestone on a public cloud journey. Few firms currently plan to replace existing technology with public cloud services, but many are experimenting with such services for low-risk use cases by complementing their existing technology stack to avoid jeopardizing the safety and reliability of existing systems.
  • Protecting existing investments in on-premises infrastructure. Many firms have already made significant investments in building or leasing data centers, but also struggle to make the business case for upfront investments in building enterprisewide DR infrastructure to keep up with data growth. Firms should consider using cloud DR as a short-term strategic approach.
  • Providing options to fail over services to another cloud provider. There were multiple high-profile public cloud outages in 2013; such outages can hinder your ability to deliver services to customers. Entire regions and cloud environments can fail; you must ensure that workarounds are in place to fail over services to another cloud. Forrester clients are also concerned about lock-in with the public cloud vendors; they’re using technologies like Virtual Private Connect to replicate critical application data that needs protection.

Companies Forrester spoke with mentioned several potential challenges:

  • Failback will be manual, cumbersome, and time-consuming. Cloud DR makes it possible to fail over and scale up compute resources quickly. But significant improvement is required in the way failback is done, especially after the DR test. Not only must the DR tools get better, but firms must make cultural and process changes in their failover and failback drills. Also most cloud DR vendors today do not provide cost-effective short-term scale-up or scale-down services.
  • Two sets of tool licenses are needed to manage the heterogeneous environment. If you’re using tools for things like automation, capacity monitoring, on-demand provisioning, or departmentwide billing, note that all of the cost savings of using public cloud for spinning up compute resources as needed will get eaten up by tool licensing if you have to manage two heterogeneous environments. It’s equally important for vendors to provide pay-per-use licenses for managing heterogeneous environments.

So why have this debate now? Interest is growing and there are challenges that DR-as-a-service vendors must address. I expect public cloud DR to gain widespread adoption over the next two years due to network bandwidth and latency enhancements, falling network connectivity prices, and the emergence of pay-per-use licensing options.

Forrester recommends that companies seriously consider leveraging cloud DR for less critical apps with smaller transaction volumes. Avoid using the same technologies and processes when leveraging cloud DR. Rely on automation for provisioning, monitoring, failover, and failback DR services. Implement technologies that can help free up resources and empower the business to move applications between different SLA tiers. In most cases, this will require several iterations and close collaboration with vendors to get the right size and architecture that you can replicate to the rest of your technology infrastructure.