Got Your Back(up)
- There are helicopters overhead, there’s rioting in the streets, severe weather from global warming threatens entire regions, and a picture of Kim Kardashian seems to have broken the internet. Plus you just pasted the wrong spreadsheet into your proposal and saved it. NOW we have a problem. How good are your backups?
Wouldn’t it be nice if we could spend the rest of our lives without having to answer that question the hard way? The reality is that no matter how reliable our computers are, people make mistakes with their data. Sometimes people even deliberately destroy data, and recently we’ve seen attackers get past well-secured networks and hold data for ransom by encrypting it. It’s inevitable that at some point in time, we’re going to find out how good our backups are when we need to use them.
The moment you discover you need something restored from a backup is not a time for surprises. The time to think about whether your backups are going to rise to the occasion is now.
Most backup strategies we run across have limitations that don’t become apparent until it’s too late. Before you go spending a ton of money on infrastructure and services in a futile one-size-fits all attempt to keep downtime at bay, let’s get a handle on what we’re protecting and how much not having it available costs. You start with the users, define what constitutes their data and follow it upstream:
- the users (internet visitors to the website, the various departments, the officers, etc.)
- the data the users need (the website content, the documents, SQL databases, etc.)
- the applications needed to work with the data (the web server, Microsoft Office, SQL Server, etc.)
- the platforms that host the applications (workstations, servers, cloud providers, etc.)
This is going to get complicated fast unless you define and limit the scope of what’s being backed up. This might include policies that dictate where documents are stored. It might include eliminating the need to recover platforms and applications by migrating applications and their data to cloud service providers, if those providers offer service level agreements – but if your data isn’t being backed up in such a way that it meets your needs, you still have work to do. Speaking of platforms, hopefully your IT department has a strategy for replacing a lost or stolen computer that takes down-time into account. Sometimes it pays to have a spare on hand and a strategy for re-imaging a computer should something awful happen to its operating system and applications.
Now that we know the scope of what needs backing up, we can develop a strategy. How often you backup is the first thing to think about. A backup you make successfully is called a “recovery point.” If you back something up once a day, you’re going to stand to lose a day’s worth of work. If it takes a long time to back something up because of the way it’s being backed up, your recovery points may be limited.
How long you keep your backups and where you keep them is the next thing to think about. Hopefully you’re not keeping all of your backups in one place, and hopefully you’re keeping them long enough that they’ll be useful when you need them. Retention policies that make gradually less-frequent backups available for longer periods of time up to the space available for retention are best. Some of our clients have very specific requirements for retention imposed upon them by governing authorities or contractual obligations.
This brings us to: how much time will it take to recover something from a backup? Should an end-user should be able to recover data? If they should, enabling them to do so is your first concern. The next concern relates to how much data you’re backing up and whether your strategy can accommodate your desired recovery points and recovery times. Copying data from place to place takes time, but taking a snapshot on a virtualization platform takes almost no time at all, and if you can launch a server directly from a backup while your IT staff recovers it to its new location, better still. Are you worried that this is getting expensive? Well, it might be – but it might not be as expensive as you think, and when you add up the cost of the down-time for all the users who are idling while they wait for data, applications and their associated platforms to come back on-line, it might be easily justified or even compelled.
If you’re worried that your backups aren’t as good as they could be, you’re in good company. Very few of us have backups that work so well there’s no room for improvement. We’d love to talk to you about backups and help bring some clarity and peace-of-mind to this popular topic, so drop us a line!