It's important to prepare your company before disaster strikes. Thus, seek out real-world business examples where businesses already have a plan in place where they'll benefit from using virtualization in a disaster recovery plan. Note how they compare with other methods, as well as their benefits and limitations. Most importantly, learn how virtualization enables businesses to recover from disasters, quickly, and then translate these methods to your own plan.

The most commonly used metrics or measures in planning for disaster migration are recovery point objective (RPO) and recovery time objective (RTO)—both measure in hours and minutes. RPO measures how far the recovered data is out of synchronization with production data when a disaster occurs. RTO measures how quickly operations are restored.

When balancing the investment in disaster recovery against the risk of a disaster, there are always known and unknown concerns. Knowing where to draw the line to suspend operations is critical to the safety of your employees and potentially your business. That being said, ask yourself is it safe to resume business operations if:

• IT systems are restored with less than optimal performance?
• There is decreased failure tolerance?
• Data is partially complete?

Let's look at some of the more common recovery methods and how they compare with a recovery methodology that includes virtualization.

Backing up to tape

Backing up data to tape is the most used and understood method to save data for retrieval in the event of an outage. Files are backed up to tape on a file-by-file basis using rotating methodologies, such as:

• Full backup: All files are backed up.
• Incremental backup: Only those files that have been changed since the last backup are saved.
• Differential backup: Only those files that were changed since the last full backup are saved.

Once the backup occurs, to ensure they are kept safe, the tapes are typically stored at an offsite location.

While backing up to tape can be part of your overall business continuity plan, it shouldn't be the only method used. This method is not full-proof and doesn't take into consideration the amount of time to fully recover from tape. It is, however, effective for localized system outages, such as recovering from a bad disk drive, so it's perfectly acceptable to have this method as part of your overall business continuity plan.

Near line and online hot sites

This kind of recovery methodology assumes that there's a failover site available. The failover site needs to be equipped with network connections, physical security, power, and cooling. Basically, you need enough equipment to ensure that business operations can safely and efficiently resume in the event of a disaster.

Many businesses have invested in online sites as well as maintaining their own location. However, there are many third-party companies who sell failover sites as a disaster recovery service.

This solution enables a quick and fully efficient recovery in the event of a disaster, such as a hurricane or a flood.

This type of disaster recovery methodology is a viable approach. It doesn't require a complete system recovery like the continuous availability solution, which is up next.

Continuous availability

In this scenario, the concept is to balance workloads over many devices. Often times these devices, or platforms, are spread out over geographical areas. Each platform is provisioned such that if there's a disruption of any kind, there's spare capacity where the workload is transferred.

Provisioning platforms is the most viable business continuity strategy. A company's business operations are always online and continuous, even if there is a disruption or a disaster.

Virtualization

So if backup, hot site or continuous availability is appropriate for your business, why then should you incorporate virtualization? Because virtualization gives you real-time flexibility, security and confidence that all of your information is safe and that your business can keep operating as if all systems are normal in a time of need.

There are degrees of disaster tolerance that can be achieved through virtualization, from high availability/low disaster tolerance down to lower availability/high disaster tolerance. Many possible configurations can be deployed along this range. All of them leverage the capabilities and efficiencies of virtualization.

With virtualization, specifically with VMware VI3, high availability is inherent at several layers. Virtual machines are designed to leverage high availability capabilities in a physical server across every virtual machine on that server. Plus, virtualization doesn't lend itself to pain points associated with backup to tape, hot sites and continuous availability.

Fault-tolerant capabilities are frequently cost prohibitive for a server running a single application. However, they become cost effective when they are shared among many virtual machines.

Outages are limited to brief restarts in this environment. Furthermore, downtime and IT service disruption is minimized and the need for stand-by hardware and the installation of additional software is eliminated.

Virtualization along with Storage Area Network (SAN) and data replication provides the highest degree of protection, since information is stored and shared in multiple places.

If there's a server failure, all of the virtual machines on that server are re-launched on other virtual servers sharing in the resource pool. This high-availability approach to virtualization is the most fail-safe. However, as was mentioned earlier, virtual machines at the recovery site hosting failed over applications reduces the dependency on redundant hardware; therefore, reducing the costs. Whether your company requires high-availability or can operate just as efficiently and effectively under a different disaster recovery scenario, virtualization can cover your needs and reduce your costs.

Virtual infrastructure combined with array-based replication

When combining a virtual infrastructure with an array-based replication strategy, you're able to replicate the virtual machine to a secondary site without human intervention on any available virtual machine.

Without virtual machine support, the failover site would be required to maintain duplicate hardware and upgrades would need to occur at the failover site each time an upgrade occurred at the production site. This is cost-prohibitive for most businesses.

The hardware independence of virtual machines means that the equipment at the failover site doesn't need to match the production site hardware.

As you can see, virtualization can benefit your business continuity strategy and plan, while saving you time, money and IT resource effort. Now let's explore how a VMware server actually works.