DR to the Cloud: Using Public Cloud as Your Recovery Site
Maintaining a secondary data center for disaster recovery is expensive. Organizations must pay for space, power, cooling, hardware and staff at a site that ideally will rarely be used. For many, especially mid-sized organizations, this cost has been difficult to justify, leaving them with inadequate recovery capabilities. The public cloud offers an alternative. Rather than maintaining a second data center, organizations can replicate data to the cloud and recover there when needed.
How Cloud DR Works
Protected systems are continuously or periodically replicated to cloud storage. During normal operations, only storage is consumed, which keeps costs relatively low. When a disaster occurs, recovery orchestration launches the systems in the cloud using the replicated data. Users then connect to the recovered applications until the primary site is restored.
Financial Benefits
The principal advantage is cost. Organizations pay for compute resources only when recovery is actually needed, rather than maintaining idle hardware. Storage costs for replicated data are typically modest. Capital expenditure on secondary hardware is eliminated.
Operational Benefits
Cloud recovery sites can be located far from the primary site, protecting against regional disasters. Capacity is effectively unlimited, so recovery is not constrained by the size of a secondary facility. Testing is easier, since test environments can be created and removed as needed.
Challenges to Address
Several considerations deserve attention. Virtual machines must be converted to run on the cloud platform, which should be tested in advance. Network configuration, including addressing and connectivity for users, must be planned carefully. Recovery time depends on how quickly systems can be launched and data made available. Failback, returning operations to the primary site after the disaster, is often more complex than failover and must also be planned.
Cost Surprises
While steady-state costs are low, an extended recovery period can generate substantial compute charges. Data egress fees when returning data to the primary site can also be significant. Organizations should model these costs in advance.
Security and Compliance
Recovered systems in the cloud must meet the same security and compliance requirements as production. Identity, encryption and network controls should be configured before they are needed.
The Bottom Line
The public cloud offers an economical and flexible recovery site, particularly for organizations that cannot justify a second data center. Paying for compute only during recovery reduces costs substantially. Careful planning for conversion, networking, failback and extended recovery costs is essential. Regular testing confirms everything works. For many organizations, cloud-based disaster recovery provides protection that was previously out of reach.