When looking at the cloud as a backup alternative, IT professionals need to consider if DRaaS solves the cloud backup problem of recovering from a disaster or ransomware attack. Disaster Recovery as a Service (DRaaS) enables you to instantiate your applications as virtual machines in your cloud backup provider’s infrastructure. Without it, cloud backup storage would force you to move all your data back on-premises to recover, which will take a lot of time, something you don’t have during a disaster.
DRaaS provides excellent protection if your data center is destroyed, especially if you don’t have an additional site to support your disaster recovery efforts. However, you don’t need all your data in the cloud to take advantage of DRaaS. In a disaster, you almost always want the last known good copy, not years-old data.
DRaaS isn’t magic, however. It does not resolve all the issues with using the cloud as a backup storage target. A combination of on-site backup storage and a copy at a Managed Service Provider (MSP) is ideal for most organizations.
The Shared Data Center DRaaS Problem
DRaaS isn’t magic. You instantiate a big piece of your data center in someone else’s data center. Other organizations may be using the resources in that data center to recover from disaster or test their disaster recovery strategies. A natural disaster impacts a region, which means your cloud backup storage provider may need to service dozens of disaster declarations simultaneously, potentially straining their resources. IT needs to assure itself of a Recovered Performance Expectation SLA.
The Cost of DRaaS Problem
One way around the shared resources problem is to use one of the mega-cloud providers (Amazon, Azure, Google) for your cloud backup storage. The problem is that their storage cost versus on-premises or an MSP is significantly higher. Seldom discussed in DRaaS circles is the cost of computing resources these mega-cloud providers charge. You may be able to get your organization back in operation, but the cost may be more than it can bear.
The Transition to DRaaS Problem
Most cloud backup providers store their data on the least expensive and lowest-performing storage in their data center to keep costs down. In most cases, they use object storage and, in some cases, may even use tape storage. You can’t run production from either of these types of storage, which means before you can return to operations, either you or the provider has to move your data to something more production friendly.
There is also an environmental problem. Most data centers are running VMware to host their VMs; most cloud providers are not. Both Amazon and Azure have a VMware cloud instance, which you can transition into but the compute costs within those environments are considerable.
The Connectivity to DRaaS Problem
Another challenge that IT professionals face when using DRaaS is getting the connections made. In a full-scale disaster, most of your personnel are likely working from another location. Making sure that users can connect to the organization’s applications and data is another challenge.
The Transition out of DRaaS Problem
Most of the above problems can be planned for and practiced. The big problem and something that few organizations can plan for, or practice, is the transition back to the original data center. In most cases, because of cost or convenience, you will want to transition out of the shared DRaaS data center as quickly as possible. If the original data center and its servers are destroyed, you must move all the data back to the replacement servers. If the original data center is mostly intact, you must carefully merge updated data while it is running in the cloud.
A full recovery will require prioritization, moving one application at a time back to the original data center. While this process is occurring and can take months, you will be paying for two data centers. A partial recovery will require careful examination of what has changed while the application runs in the cloud and only restoring that data. You will likely have to run a net new complete backup on-premises to ensure your backup metadata is in sync with the actual data.
Is There a Better Way?
The DRaaS problems are a lot to overcome to move back up storage out of the data center. Remember, your backup exists because someday, you might need to recover. Cloud backup storage continually throws up challenges to fast recovery.
The primary motivation for using cloud backup storage is to eliminate the upfront cost of backup storage, free up data center footprint and eliminate scaling problems. These on-premises backup storage problems all exist because existing backup storage technology is based on over two decades old code.
StorONE has rewritten the storage IO stack, creating a new, more efficient storage engine. This engine can take full advantage of modern hardware innovations like fast SSDs and high-density HDDs without compromising performance or suffering through slow RAID rebuilds. StorONE’s new Backup Infrastructure as a Service (BIaaS) solves the final on-premises backup storage challenge, the upfront cost. BIaaS brings the cloud OPEX model on-premises while making your existing backup software better and your organization more resilient to a ransomware attack. If BIaaS is used on-premises with an off-site copy sent to an MSP, organizations have an affordable and complete recovery strategy.
In this week’s live webinar, “Beat the High Cost of Backup Storage with an On-Premises Cloud,” StorONE’s CEO Gal Naor and I will be discussing the pros and cons of using the cloud for backup storage.
[…] The move to an OPEX model should not force the organization to compromise on the intended purpose of the backup storage target. Cloud backup storage, one popular means to deliver the OPEX model, can create additional challenges, especially in large-scale recovery efforts caused by a ransomware attack or natural disaster. The cloud may not be the best place to put backup data, and Disaster Recovery as a Service (DRaaS) does not overcome all cloud recovery challenges. […]