For many small to mid-sized companies an initial data protection strategy often consists of utilities that come with the operating system or environment or one of the "default" data protection products for those platforms. As these companies grow so do their on-premise data protection demands, and their expectations. Often the original data protection solutions are inadequate and they look for a new application. As part of this search they include cloud focused products that can also aid in off-site data movement and disaster recovery (DR). The challenge for the small and mid-sized company is finding a solution that addresses both their on-premise data protection and off-premise DR needs.

Mid-virtualization Data Protection Challenge

   

One of the first expectations to grow as the smaller company grows is a demand for faster recovery of failed servers. Simple data protection may seem adequate at first but most soon realize that server recovery is one of the primary objectives for their data protection efforts in the first place. In this regard, small and mid-sized companies have time pressures to bring applications back on line, just like large enterprise data centers do. Server virtualization is well suited to this goal and fast application recovery is often the initial motivation to begin a virtualization project.

The challenge with leveraging virtualization as a rapid recovery tool is that even mid-sized companies typically can’t afford to go “all in” with this technology. For them, like most businesses, virtualization is an incremental process not a one-time event. As this virtualization process unfolds the ideal first step is to set up a physical host and use that host as a target for migrating their current stand alone servers. In this way the host can act as a stand-by replacement for any failed application server. This migration is a form of backup not a conversion to virtualization and they will only run the migrated image if the physical server fails.

While this seems like an ideal first step to virtualization most small to mid-sized companies are challenged with how to get an updated image to the host in an ongoing fashion so that the host is kept current with the stand alone server. As a result a VM specific replication or backup application needs to be purchased since the free migration utilities available are typically one-time, one-way options. Also, this image can’t easily be used for single file restorations, so legacy backups still need to be run. Finally, there is the cost of a physical host that basically sits idle until a stand alone server fails.

There are also operational problems that occur as the environment continues to grow and moves to the next step in the virtualization process, embracing virtualization to eliminate server sprawl. Essentially new VMs are created on the host instead of a new stand alone server being purchased. This means the host now contains unique data, not just backup images of stand alone systems.

Now the small or mid-sized company needs an application that can protect the virtual machines containing this unique data. Also, in order to maintain rapid recovery and support the virtual machines in the event of a host failure, they may need an alternate host as well. In reality many companies don't need more than one physical host and are forced to buy a second host only for redundancy. Once again cost becomes a key challenge, the cost of virtualization-specific data protection software and the cost of a secondary host. And a separate data protection application will still be needed for stand alone servers that are not yet virtualized.

An ideal solution to the virtualization-as-a-process data protection problem is to use a backup appliance that has backup storage and software to provide data protection to both physical and virtual servers. Companies like Quorum have created appliances that not only provide the software and hardware but can also virtualize server instances on the appliance.

The virtualization component on the appliance means that the basics of on-premise data protection and restoration can be addressed, along with faster recovery. It also eliminates the need for a stand-by host in case of a primary host failure. If a stand alone physical server fails then the virtualized images of those physical servers can be launched directly from the backup appliance. Similarly, if a host supporting multiple VMs fails, those virtual machines can be restarted and run until repairs are made to the fallen hardware. Not only does this provide ideal return to operations capabilities it also lowers costs. Instead of buying stand alone “DR” hosts, the backup appliance can be used.

Leveraging The Cloud

Another expectation that grows as the small company grows is the ability to recover in case of a site disaster. This is the off-premise challenge that the IT team needs to address. For small and mid-sized companies, setting up a secondary data center is not only cost-prohibitive but also more complex than their IT personnel can typically handle, often involving dedicated hardware, special scripting, etc. For many a common solution to investigate is the cloud.

But the cloud brings its own set of challenges. First, as described above, unless a local appliance is used then speedy recoveries becomes a significant issue since all recoveries from the cloud have to come across an internet connection. While this may be acceptable for single file recovery, it’s not appropriate for full-server or even smaller database recoveries.

Even with a hybrid backup appliance recoveries can be problematic in the DR context where company survival may depend on how fast applications can be back on line. When the primary site is lost the hybrid appliance in that site is lost as well. While data is secure in the cloud, it all needs to be recovered through the cloud. This means a secondary site has to be established, something that not every small or mid-sized company has, and a new appliance installed before recoveries can begin. And once again, all those server images need to be transferred across an internet connection.

Quorum’s Hybrid Cloud Solution, as an example, can address this problem in two ways. First, for companies with two locations, a second appliance can be installed and have data replicated to it. In this way it has the most recent copy of server images almost as soon as the primary site does, so when a disaster strikes the data is already in place at the secondary site. The hybrid approach has an advantage in that the recovery appliance is the ONLY device that needs to be in the remote location since servers and virtual machines can be restarted on it directly.

For companies without a suitable secondary site this hybrid cloud model can be extended to a full cloud recovery model. As Storage Switzerland discussed in the recent article “Does Real DR Protection Require a DR Site?” some hybrid cloud appliances can also provide cloud DR. Cloud DR allows for companies without a second site, or  those that simply don’t have the resources to manage a second site, to leverage the cloud compute capabilities of a provider like Quorum.

Conclusion

The “hybrid” part of Hybrid Cloud DR is important. The ability to provide high quality, on-premise backup and rapid recovery is critical. It’s also important that an on-premise functionality support, and even promote, the move to virtualization. Finally, the cloud part of the solution is important as well. But when the cloud is used, recovery time must be kept in mind and leveraging either a second appliance or the cloud provider itself is an ideal solution.

Previous Entry: “Rapid Recovery-Part of The Next Generation of Data Protection”