Cloud backup is an example of the increasing convergence of backup and disaster recovery. Restoring large amounts of data over a WAN can be very slow, but disaster recovery as a service has emerged as the answer to the cloud failover issues many organizations encounter.
Today, DRaaS means the instantiation of an organization's applications as virtual machines in a provider's cloud. It initially resolves cloud failover issues by not moving data at all. Instead, applications use cloud compute resources while data is accessed from cloud storage copies. DRaaS can be especially attractive to organizations that do not have a secondary site with an embedded IT staff.
DRaaS may appear to be an ideal solution at first glance, but there are three questions IT planners need to have answered before using the technology in their cloud strategy.
How long will it take an application to be available once a disaster has been declared?
Many providers use the term instant to describe DRaaS recovery. In most cases, application access time can be as much as two hours.
What will application performance be like in the provider's cloud?
This question is especially important during a regional issue where dozens of businesses may declare a disaster at the same time. Users and application owners expect application recovery to perform similar to that of production. The scalability of a provider's compute infrastructure and the type of storage used to recover applications are critical factors in delivering acceptable performance during a cloud failover situation.
How will applications return to service once the disaster has passed?
Most, if not all, DRaaS providers do not want the application to reside in the cloud permanently, so a planned failback strategy is vital.
DRaaS products have the additional capability to perform a slow replication back to the primary site over the WAN while running the organization's application in their cloud and then doing a quick sync after completing the replication.
The only problems with this approach are whether an organization is comfortable with running in the cloud for that long, and if it can afford to run applications in the provider's cloud for an extended period while replication happens. DRaaS can be expensive because you pay for the cloud compute resources used to run applications in the cloud in addition to your typical cloud storage costs.
The integration of disaster recovery with backup should substantially reduce the cost of both processes and the time required to manage them. Centralized processes should also eliminate the human errors associated with missing backup or replication jobs.
For IT professionals, the goal of cloud failover and recovery is to simplify data movement as much as possible while meeting application performance in its recovered state. Outlining how to return the application to the primary data center after the disaster has passed, and integrating backup and replication, goes a long way to achieving those goals.
Backup products evolve to include DR
With most cloud recovery, plan on failing workloads back ASAP
Understand the services DRaaS providers offer