Best practices, tips and tools for VMware virtual recovery and backup
A comprehensive collection of articles, videos and more, hand-picked by our editors
Cloud-based disaster recovery goes by many names and is offered by many public cloud providers. Protecting your compute environment by hosting it in the cloud can dramatically lower costs and simplify DR planning.
Many IT shops rely on tape-based backups as part of their disaster recovery (DR) strategy, but forget about where IT will restore these backups during a complete data center disaster. This major planning gap usually boils down to the prohibitive cost of operating a complete second infrastructure and the failure of the IT shop to prioritize DR planning.
Businesses with a physical DR operating environment must either build and maintain a second data center for it or negotiate leased space in a colocation facility. VMware shops running Site Recovery Manager (SRM) benefit from fully automated site recovery and migration in the event of a data center disaster, but the physical backup data center is still costly -- and potentially poorly managed.
One DR choice can alleviate these concerns: cloud. It's not necessarily a cure-all, but can offer many benefits to underprepared businesses.
Hosting disaster recovery in the cloud has upsides
When an organization chooses to build a cloud-based DR strategy, it will have an easier time moving data to the cloud than moving IT hardware to a new location. The task is as easy as negotiating with a new cloud provider and migrating the data.
If a public cloud provider hosts your disaster recovery environment, you can drop the acquisition and maintenance costs for redundant server, network and storage hardware. This strategy avoids negotiating storage array prices, refreshing servers every three to five years and making calls to support for problems, such as a bad switch port. It also relieves the overhead of maintaining firmware and renewing support for each device. In some cases, you could bypass the additional work of software licensing.
In the event of a regional disaster, cloud-based DR enables an organization to restore in a different part of the country or world. A company headquartered in Florida may want to have their data center operations close by, but, when a hurricane strikes, the ability to restore virtual machines (VMs) in a cloud hosted out of Iowa could mean continuing IT business as usual.
The downside to cloud-based disaster recovery
Of course, there are a few disadvantages to cloud-based DR. Security is one of the main hesitations for moving DR to public cloud. If a business doesn't trust a public cloud provider to run its production workloads, should it trust the same cloud provider with the same workloads in a disaster situation? Possibly; decision-makers may perceive the risk of a public cloud environment as acceptable during a temporary emergency.
When VMs have very low recovery point objective and recovery time objective (RPO/RTO) requirements, businesses should incorporate VMware Site Recovery Manager, or a similar product, into their disaster recovery plans. That added automation tool can speed data recovery times over manual processes, so you meet those RPO/RTO goals. The replication features built into SRM enable Disaster Recovery as a Service, copying changed blocks of VM data without storage lock-in.
VMware vCloud Director enables IT shops to create virtual data centers quickly and easily. SRM 5.1 and VMware vCloud Director 5.1 have very little integration with one another, however. SRM can be used to fail VMs into the virtual environment, but won't add them into vCloud Director. If you're handy enough at scripting, you can make SRM and vCloud Director cooperate, but they will not directly interact on their own. This makes failing VMs into the vCloud Director environment via SRM a very limited option. However, some cloud service providers are offering DR to the cloud services, using VMware vSphere 5 and SRM.
Not many IT infrastructures are 100% virtualized, so physical machines will be part of most organizations' disaster plans. Cloud environments rely heavily on virtualization, so they have very little to offer the physical-machine victims of a site failure. If the cloud provider has a traditional colocation business, additional planning -- and possibly additional contracts -- may secure protection for physical servers.
While many organizations do not consider public cloud a viable production infrastructure option today, many are happily considering a cloud-based disaster recovery scenario. The cost savings and ease of implementation make cloud attractive to businesses of all sizes.