Originally data protection was focused on backup windows and getting backups done before users came to work in the morning. And disaster recovery (DR) used to mean getting those backups (initially tapes, but now disk backups as well) stored in a safe place off-site. But as the overall speed of business picks up and more companies rely on their computer systems to run their businesses, they’re starting to realize the cost of downtime. Recovery therefore becomes the focus, and the need for a ‘real’ DR solution is recognized by more of these companies, including those in the mid-market space and smaller.

Historically, the ways companies addressed their DR requirements have run the gamut from something as elaborate as an off-site data center with applications loaded on hot-spare systems to something as simple as a server closet in a small sales office running a remote backup appliance. While the range of DR solutions is broad, the common denominator has always been the remote site. Now, technologies like the cloud and server virtualization have brought new alternatives to the table and companies are asking if they even need a DR site in order to get a real DR solution.

What is ‘real DR’?

Similar to the expression “backup is really all about restore”, disaster recovery is really all about the resumption of services. This means that getting data back from a secure location is just the first step in recovering from a downtime situation. Critical applications must be restarted and users brought back online. For traditional IT environments this means having hardware and software available to run critical servers and a solid process by which IT can reload applications and data sets in order to restart services within an acceptable time frame.

For this reason DR often means a ‘big’ solution - big infrastructure, big implementation project and big ongoing costs - plus a remote site. Sending backups to the cloud keeps data safe, but doesn’t mean it can be brought back quickly, and more importantly, it doesn’t guarantee that services will be restarted in time. But a new kind of DR solution that leverages on-site ‘hybrid’ cloud appliances and server virtualization may address that problem.

Most companies are on a path to virtualizing some portion (often a large portion) of their physical server infrastructures. These virtual machine images can be backed up to a storage/compute appliance, even in a remote location, and kept updated so that they’re available in the event that the primary server goes down. When this happens, the remote VM can be started and users can be switched over to the secondary appliance, achieving recovery in a short period of time. Now this technology has been combined with the cloud to replace the need for a dedicated remote facility and provide an effective DR solution.

Real DR in the Cloud

Products like Quorum’s Hybrid Cloud DR solution are set up locally to capture backups of virtual machine images, or to perform a physical to virtual conversion of physical servers. Then these virtual versions of all of a company’s servers are replicated to the cloud on the backend, similar to the way cloud backup services have operated. But users can actually run these VMs from the cloud on demand, providing a very short recovery time - a ‘real DR’ solution.

The appliance stores copies of VMs that are replicated to the cloud, providing a local recovery point that can be used when a single server goes down, or a simple stand-in to use when primary servers undergo maintenance. Updating VMs in the cloud incrementally saves bandwidth and assures that they’re current and ready to run when needed.

DR Testing

Disaster recovery protection is essentially insurance and users must have confidence that it will perform when required, so real DR systems must be tested regularly. Unfortunately, testing is something that’s not always done or at least not done frequently enough to be of value. Neglecting to maintain system changes in the standby infrastructure as well as overall time and cost constraints make testing a popular task to put off.

But hybrid cloud DR systems are simple to test. Updating VM images in the cloud is done as part of the regular backup process so systems are always current. And, restarting VMs from their cloud location is also easy, enabling IT to run tests as often as desired and improving the confidence level of IT and of executive management. With the Quorum solution for example, all VMs of a company’s servers are auto-tested daily, ensuring their DR plan actually works. Customers can also run tests on demand with a click of a button, as often as every 15 minutes if desired.

Recovery Data Center-as-a-Service

This is an ideal use case for the benefits of the cloud since DR is very much a ‘part time’ requirement, at least the recovery portion. Having dedicated infrastructure tied up waiting for a downtime event is a very inefficient use of resources. Server virtualization is the other technology that makes this work, allowing users to protect their data at the VM level and share compute resources at the cloud site. This enables the asset utilization that drives the economics and makes this cloud solution more appealing.

Conclusion

Effective disaster recovery doesn’t just mean getting data off-site, or getting data back, it includes restarting critical applications and bringing users back online. Historically, this kind of DR protection has required a dedicated location with the infrastructure available to run these applications. It’s required a DR site. Leveraging the cloud and server virtualization a new technology has emerged which is changing this. Solutions like Quorum’s Hybrid Cloud DR solution can provide a local VM backup repository and cloud gateway providing the ability to run critical VMs in the cloud, giving companies real DR protection without a DR site.

Quorum is a client of Storage Switzerland 

Previous Entry: “What is Big Data Analytics”