I hear a lot of folks asking for cloud solutions. Many think cloud solutions are cheaper, faster and more secure than alternative means to store and maintain their data. Just a note, if you’re interested in learning about why or why not the cloud is a good solution for your organization, consider our recent discussion on the myths about cloud.
In any event, most people think that cloud is a set it and forget it technology. But what most people forget is figuring out how to backup data in the cloud and testing whether that data is recoverable in case you need it.
How can you backup your cloud data?
There are several options when it comes to figuring out a cloud backup strategy that works for you. Here are a couple common options:
Storing it in one cloud— One common approach is to back up your applications and data with the same cloud solution that its already hosted. While this approach is easier and cheaper than implementing alternative backup methods, it lacks isolation—especially if your data and application backups are not separated in a different part of the cloud (people in the industry call this partitioning).
The problem with keeping your backups in the same storage as your production data and applications is that in the event of an outage or security breach, your entire system could fail, leaving you without a backup to recover from.
Taking a hybrid approach—many organizations have started to take a hybrid approach when it comes to backing up their cloud data. What many folks are doing is backup up cloud-hosted applications and data locally, where you have full control over data retention, storage and security. This provides you with a higher degree of isolation from your production data sources and applications and minimizes the effects of a cloud failure.
The downside with this method is it can introduce slowness issues and delay restores because of physical separation between the off-premises cloud and the on-site backup environments.
Cloud-to-cloud backups—another alternative that some experts are advising is to back up your data in a separate cloud. This approach allows for isolation, like having the hybrid cloud solution, and also can provide quicker recovery times.
The downside to cloud-to-cloud solutions is they are often the most expensive option, since providers will bill for having increased bandwidth usage across your connections. In addition, use of different infrastructure needs such as an inter-cloud VPN may increase your costs.
My advice: consult an expert on cloud solutions before signing on the dotted line. Make sure you have a strategy for backing up your data that will fit your budget and expectations in the event you need to recover data quickly.
What tools can support your cloud backup strategy?
Nowadays, there are tons of vendors that say they do cloud. Even Google and Amazon offer a variety of storage tools. The problem with a lot of cloud tools—especially the ones that are free—come with unexpected consequences.
For instance, a couple of years ago, when Dropbox stopped working for well over 2 days, many businesses that depended on Dropbox were unable to work. The reason? While Dropbox was interested in getting their services back up and running, your data was no more important to them than the personal account you might store family photos in. I hate when organizations face the true consequences of free products, ending up costing them ten times or more the cost of maintaining a cloud solution that is more controllable.
Whatever cloud solution you determine is the best strategy for backups, remember that practice makes perfect. That means regular testing of your cloud backups and your disaster recovery process to make sure your recovery runs smoothly in the event of a real disaster. Most experts recommend testing at least quarterly—especially as cloud environments may get upgraded or changed (by the vendor) over short periods of time.
How can you make sure your cloud backup strategy has failover protection?
Failover and replication are two critical parts of a cloud backup solution. For failover, most experts on cloud recommend that you choose a solution that has the capability to failover to a different location in case one facility goes down.
This likely would mean having a cloud solution that has hosting in two geographic areas. The trade off to having failover protection across geography is that you will probably notice slowness in accessing data the farther the data centers are from each other. In essence, you experience latency when replicating data to a data center in a different location.
When you replicate work or data, make sure to define a specific sequence of events to avoid having bandwidth issues at your office. Create a schedule to prioritize critical data backups so that you ensure the most important systems are working immediately in the event a failover event occurs (i.e., you need to pull backed up data from a different location to restore systems at your office.
Not sure whether your backup strategy will actually recover your data? Consider a network assessment to see if your cloud solution is working the way you expect.