Cloud-based data protection is appealing and many IT planners are considering it as an option. Can on-premises data protection infrastructures still compete? An on-premises infrastructure still has some technical competitive advantages and with consumption based IT the cloud business model advantage can be neutralized.
The big advantage that an on-premises infrastructure has is latency. The backup software does not need to go across a relatively thin Internet connection to backup or restore data. An on-premises infrastructure doesn’t have to worry about initial seeding times or how to do a large restore.
An on-premises backup infrastructure is also in a better position to deal with rapid recoveries. While cloud-based backups can leverage the same boot from backup capabilities as an on-premises solution that recovery needs to occur in the cloud. Recovery of even a single server means that the networking infrastructure needs to be reconfigured, and while some of the reconfiguration can occur automatically there is almost always some manual intervention needed. Additionally, many organizations don’t have the in-house networking skill sets to orchestrate this type of fail-over.
None of the networking issues exist in an on-premises boot from backup (a.k.a. Instant recovery). The users are already inside the network as is the recovered server.
Another challenge with a cloud based boot from backup recovery is that at some point the organization will want to move that system out of the cloud and back on-premises. Cloud-based backup vendors will say that the customer can just reverse replicate the data and fail-back when replication is complete. What if that server is associated with dozens of terabytes of data? The replication will take a very long time and the organization may incur additional charges from the provider while they are running in their cloud. They may also incur significant egress charges depending on the amount of data they retrieve from the cloud. Additionally, there is also a concern of a network failure; does the replication job have to restart? In many cases it does. Again, these issues go away in an on-premises scenario.
The cloud-based backup vendors often focus on their disaster recovery advantage because the customer’s data is already off-site and they don’t have to pay to equip a complete disaster recovery site. The reality is that most organizations rarely if ever experience a data center destroying disaster. Instead it is much more common to have a single server fail, and in that case the rapid recovery advantage of on-premises solutions shines through.
Additionally, several on-premises data protection solutions can replicate their backups to the cloud, and initiate cloud based recoveries. In this case the customer has the best of both worlds; on-premises recovery for the most common “disaster”, server or storage system failure, and the cloud for a data center failure.
It all comes down to the business model. If the organization likes the financial all-OPEX model of the cloud then it seems like the on-premises solutions lose. Now though companies like HPE, with its GreenLake service, provide consumption based IT which is similar but in some ways superior to the cloud model because it automatically scales down for resources that are not in use. Most cloud vendors report that their customers never scale down their storage consumption (even though they can) or their compute resources utilization.
HPE partners with Veeam to provide data protection for GreenLake customers. To learn more about getting the best of both worlds for your organization watch our on-demand webinar with HPE, Veeam and Storage Switzerland “How Consumption Based Data Protection can Provide Peace of Mind”.