May 24, 2018
Whether it be ransomware, a natural disaster, user error or simple hardware failure, backups are an absolute must in today’s IT world. Yet, a surprising number of companies wrongly think that something like mirrored drives are enough, or someone remembering to plug in a USB drive each day will save the day. Nope.
When we set out to design a backup system that was resilient to all forms of failure, some said that there was no such thing as a perfect backup. Not only have we created, what we believe to be, the best backup, but through some simple design choices, we’ve limited the cost pretty significantly.
There’s a old sayings in backup called the “3-2-1 backup” or the “Backup Rule of 3”. Basically, it states: Have at least three copies of your data. Store the copies on two different media. Keep one backup copy offsite.
Interpretation of this rule varies though. One example of this would be having the copy on your computer or server (that’s one copy), backing it up to a tape (that’s the different media plus a copy), and then uploading a copy to a cloud storage provider (third copy, also possibly a different media). Another example could be: a copy on your computer, a copy on a USB flash disk, and a copy at your friend’s house. Technically, both of these examples meet the 3-2-1 rule, but are clearly dramatically different in robustness, reliability and cost. So, does the 3-2-1 rule even work any more?
Look, having a good backup is very important. Something bad will happen to your data at some point and if that data runs your business, guess what. Data gone = out of business. That may sound like an exaggeration, and it would be if we hadn’t seen it happen before. We dread that panicked call from a business owner who got ransomware, the IT guy couldn’t help, and now she’s looking for someone to help get them back up and running.
In this scenario, if a business has good backups, it is just a matter of rolling back and continuing business as usual. If not, sorry, but you’re out of luck.
So, what makes a “good” backup. It isn’t as simple as a rule, nor something businesses should handle on their own. Put simply, a “good” backup strategy should:
- Be 100% automated. No human intervention necessary for routine backups.
- Be tested by automated means daily. That usually means automatically restoring the backup to a virtual machine and running some automated tests.
- Be tested manually for a full disaster recovery scenario once per quarter.
- Allow you to resume business operations in less than a day in the case of a total site failure. A total site failure means loss of physical hardware, a major virus attack, or natural disaster.
- Keep your data secure, or at least not be the primary attack vector. In other words, your backup should be just as secure as your data on site.
- Be well documented.
We have yet to run into a backup system, not created by us, that meets all these criteria. So, how does our backup work? We don’t want to give away trade secrets, but basically a system for on premise servers works like this:
We install an appliance running our hardware and software. Your servers back up to our appliance using various tools, from ones as simple as built in Windows backup, to great tools such as Veeam for more complex environments. That appliance replicates to our data center where data is stored geographically near to our clients. This is also where we test the automated restore to virtual machines, boot up those machines and run tests on them. The data is also replicated to two different cloud storage providers for added redundancy.
Using the above method, we can restore a few files file very quickly from the appliance to your server. We can spin up a copy of your entire on premise data center in ours in case a natural disaster hits your building. If a major event hits both the client site and our data center, the data is further protected in globally redundant cloud storage.
We’re pretty proud of our backup. Not only is to cost effective for backup, but especially so when considering what a day or two of business downtime could cost. In addition, our clients sleep better and we do too, knowing that when the inevitable happens, the fix is just a few keystrokes.
How’s your backup work?