Regarding backup and disaster recovery, the cloud is ideal for getting data off-site. It is also ideal for disaster recovery thanks to the many Disaster Recovery as a Service (DRaaS) offerings on the market. The cloud is appealing because for many organizations, even large ones, the cost of maintaining a secondary site gets expensive. The cloud provides this secondary site at a very competitive rate. Now with DRaaS, data centers can leverage their cloud backup service or a replication service to instantiate applications in the provider’s cloud. The problem is that many organizations let their cloud backup also provide long-term data retention, replacing traditional archive, which raises cloud storage costs considerably. To maximize cloud capabilities, a new strategy is needed.
Hybrid Cloud Backup 1.0
Most cloud backup solutions leverage an appliance on-site that stores the initial backup locally and then replicates all of the backup data to the cloud. The appliance can be used to meet the most common recovery requests locally without incurring cloud latency. This is commonly called hybrid cloud backup and gets around one of the cloud’s biggest problems; the latency of the internet connection.
Hybrid Cloud Backup 2.0
The problem with Hybrid Cloud Backup 1.0 is that the cloud becomes the historical repository for all backups. Unfortunately, this exacerbates another cloud challenge, the cost of long-term storage. As Storage Switzerland stated in its column, “What’s better than Cloud Storage for Cold Data?“, the cost of using the cloud for long term data retention is expensive. Those periodic costs add up and after 5 to 10 years of making the payments, many data centers find they can provide an on-premises archive for long-term data retention at considerably cheaper costs.
As a result, a new hybrid cloud backup model is needed: Hybrid Cloud Backup 2.0. This new hybrid cloud backup model builds on the current one. Data centers should continue to use an on-premises appliance to backup data locally and replicate it to the cloud for disaster recovery purposes. The difference in Hybrid Cloud Backup 2.0 and previous implementations is that the backup process should only store the most recent backups. With most cloud backup solutions, the number of backup copies can be easily reduced by setting the retention times on backups to store one or two full backup copies and two or three weeks of incremental or differential backup copies.
While most IT professionals will agree that the majority of their restores are from the most recent backup data set, there is a valid concern with such a short-term retention schedule; data loss. In reality, the only data that can be lost in the short retention scenario is data that is deleted or versions of files that are overwritten. As the full and incremental backups age out, deleted and previous versions of files will eventually be lost. For obvious reasons, this data needs to be saved. The big questions regarding the new hybrid backup model are what will save that data and where will it be saved?
Hybrid Cloud Backup 2.0 stores deleted and previous versions of files in an on-premises archive, one that mimics cloud capabilities, so time spent managing it and costs to maintain it are kept to a minimum. It should also provide the option to support tape to both keep costs down and create a disaster recovery copy of the data it retains. The on-premises archive should leverage the cloud-in-the-box concept as we discussed in our article “Cloud in a Box – The On-site Cloud“.
Since the cloud-in-a-box solution is accessible via an NFS, CIFS or Object Storage mount point, there is a variety of ways to get data to it. First, the backup application could “restore” data from the incremental backup copies into the on-premises cloud. This would populate the archive with native files that it could store and manage. Data could also simply be copied to it from the primary storage systems or applications. Eventually, production data that is not changing could be moved to the cloud-in-a-box, thinning the size of the full backup and further reducing the amount of capacity required in the cloud.
The Impact of Hybrid Cloud 2.0
A Hybrid Cloud 2.0 strategy would not hurt cloud providers. While giving up some of the recurring profits associated with recurring storage costs, they would gain more and larger customers. Data Centers that thought the cloud was too expensive to store their entire backup footprint can now cost justify its use since that footprint is demonstrably smaller and that footprint will grow at a much smaller rate. These data centers can leverage the cloud for its strengths, establish the second site, and avoid its weakness, long-term costs.
Sponsored by Fujifilm Dternity, Powered by StrongBox