What should an AWS cloud disaster recovery plan contain?

Organizations that depend completely on public cloud providers, such as Amazon Web Services, need to create and...

implement a comprehensive cloud disaster recovery plan for mission-critical workloads. If not, service disruptions from outages can result in user dissatisfaction and potential financial loss. There are familiar tactics that enterprise IT can use to help improve disaster recovery for public cloud.

Often, problems arise because organizations assume a level of provider reliability that doesn't exist. Businesses sometimes underestimate or forget the effect a few hours of downtime can have -- and never adequately prepare a disaster recovery (DR) plan. Take the time to identify mission-critical applications that reside on premises or in the cloud, and ask what happens to the business if those applications go down.

Don't ignore the importance of testing after implementing a cloud disaster recovery plan. Many organizations never check to ensure that the DR strategy works. Implement a regular testing regime to ensure that failover and failback activities work and can adequately handle the user load. In-house testing scenarios include user error recovery to retrieve lost data, lost connectivity to a site and power loss to a site.

Organizations that fully depend on the public cloud may be limited in testing scenarios, but they can still invoke traffic failover to gauge AWS Auto Scaling response and performance. Testing is also a perfect time to work with any monitoring and alerting tools, as well as verify that personnel are adequately informed, and tickets are generated and addressed as expected.

Finally, treat DR as an ongoing process rather than a static, one-time project. Revisit and re-evaluate the DR strategy several times per year. This is an ideal time to measure DR and any response to inevitable changes taking place across the business and the public cloud. For instance, a cloud provider may offer new services, regions or availability zones. All of these can offer opportunities to improve a cloud disaster recovery plan while staying within an established budget. In addition, compliance requirements may change, so review which AWS region is best.

No application is 100% available 100% of the time. Public cloud and telecom providers experience occasional disruptions, resulting in potential service disruptions. Every organization with workloads in the public cloud must expect availability problems and implement a suitable cloud disaster recovery plan to remain flexible and respond accordingly.