Most disaster recovery tools available today make heavy use of disaster recovery orchestration. A form of automating DR, orchestration can be tremendously beneficial in a disaster recovery situation, but there are some disadvantages to its use.
There are two main benefits to disaster recovery orchestration. The first benefit is that orchestration greatly reduces complexity. Imagine that a virtual machine (VM) running in your data center needs to fail over to the cloud. There are a lot of things that have to happen in order for the failover to succeed. For instance, the VM may need to be converted from Microsoft Hyper-V to VMware, or vice versa.
Likewise, the VM will need to be assigned a new IP address, and domain name system records referencing the VM will need to be updated. These are just a few of the tasks that admins commonly perform during a failover. Orchestration keeps an administrator from having to perform these tasks manually. This helps to prevent human error and enables the failover to occur quicker.
The other significant benefit to using orchestration for DR is that it can help with scale. Given the tasks that admins must perform during a manual failover, just imagine how time-consuming it would be to manually fail over 1,000 virtual machines. Orchestration makes bulk DR failover and failback practical.
As helpful as disaster recovery orchestration may be, it does have its drawbacks. The most significant disadvantage to using orchestration is that it does not always work the way it is supposed to. Few tools are infallible, but when it comes to DR, the consequences of failure are significant.
While it is true that an organization can use disaster recovery testing to help ensure that orchestration works the way that it is supposed to, things can still go wrong. Making a small, seemingly insignificant change to a virtualization infrastructure can cause an automated function to break.
When orchestration fails, the organization's VMs may be left in an unknown, partially migrated state. It would then be up to the IT staff to figure out how far the failover had progressed and what they need to do to return the VMs to a functional state. Depending on the nature of the problem, this can end up being far more work than if the organization had simply opted to perform a manual failover.
The benefits of disaster recovery orchestration usually far outweigh the potential for problems. Because the potential for problems does exist, however, it is critically important to perform disaster recovery testing on a frequent basis.