There is backup and then there is backup. To meet the backup and recovery needs of today’s organizations, they need to verify that the selected backup appliance includes the features needed to protect their environment today and positions them to meet their needs into the foreseeable future. In this third installment of DCIG’s interview with STORServer President Bill Smoldt, he describes the new must-have features that backup appliances must offer.
Jerome: What do you consider the new must-have features of backup appliances?
Bill: The two big features that jump right out would be deduplication and replication. These go hand in hand. STORServer can replicate between appliances or replicate to a public cloud or private cloud without doing deduplication.
However, it makes a lot more sense with deduplication where appliances have like deduplicating algorithms at both sites and they only have to send data that has not been sent before in some duplicate fashion. Using deduplication, we can detect a duplicate chunk of data and just send a pointer into a hashing table instead of sending the data.
Those are features that customers certainly demand at this point. While they are not always necessities, they are certainly on their checklists so we built them into the appliances. We will see a lot more deduplication in different places in the future because it does permit our appliance to store much more data and we have more of that feature moved down further down into our devices.
Instant recovery is another one of these features that is becoming more important to many customers. This takes two forms. One, they want to bring up a machine while it is still in our backup appliances, such as a virtual machine. Alternatively, they want to mount a disk or a volume right out of our appliance so they can browse the files on the device or use the device. That is another feature that customers are demanding at this point.
Jerome: What steps is STORServer taking to accommodate public cloud storage connectivity?
Bill: It is absolutely critical that STORServer offers that service. We offer it on our own cloud that we run. But we also have managed service providers (MSPs) that have bought appliances from us and who offer cloud services. That permits us to exchange data with those MSPs to resell their services and connect our customers with them, and then they also resell our appliances.
That part gets really interesting from an MSP perspective because they can also offer a complete disaster recovery (DR) service within their own cloud and help a customer run their data center inside their cloud. They really like that aspect of DR. So if you lose a facility, you can be back up and running with the MSP more quickly than if you had to try and rebuild your own environment in your data center.
In regards to cloud adoption, most of the customers we deal with directly are reticent to send data to a public cloud at this point. Part of that has to do with some of the things that have gone on with some of the cloud providers. But others are adopting it just fine.
We see a real adoption rate at the low end, where perhaps a customer does not have enough data that requires an appliance at their site and they are able to depend on a backup straight to the cloud. That is what our MSPs typically offer. We have certainly seen it in the high end where end users can afford the bandwidth that they need to replicate all of their data.
Yet, what is enabling much of the cloud connectivity is the continuing drop in communication costs coupled with its increasing speeds. For example, SMB sites now have communication speeds to the Internet that are equivalent to what local area networks (LANs) had 10 years ago. That is making a big difference.
From strictly a restore perspective, the restore speeds that a customer typically needs to restore data to their own environment is still too slow, so they often need an appliance at their own site. In these cases, we do what we call disaster recovery to the cloud.
There are several layers here. There is backup as a service, disaster recovery as a service, and a whole infrastructure as a service where part of the DR infrastructure is restored into a cloud facility.
STORServer can offer all those different levels at different prices. Most of the adoption rate has happened at a different level where customers want to create their own hybrid cloud or even a private cloud of their own.
Quite often, we will have one of our existing appliance customers during a technology refresh put the new appliance on their own site and then move their old appliance off to a remote site as a target for replication. Or, they have a site that perhaps has a few computers and they can do their own recovery. Most of our customers are more comfortable with that because they still own the data and it is never out of their control.
Many still exhibit a bit of hesitation when putting data out in the public cloud, but that is likely going to change fairly quickly. Public cloud storage is going to get less expensive and that is happening rapidly. As that public cloud storage gets less expensive, organizations will put more and more data there, particularly for long-term archive. That just makes perfect sense plus it provides a DR layer.
Given fast enough communication speeds, we will see more backup as a service in the cloud. We see that in our own customer base as they go from 100 megabits to 1 gigabit to, in some cases, 10 gigabits infrastructure speeds between sites. Then it makes more sense to provide centralized backup and recovery too.