Personal Backup Strategies
Given that a fair amount of my work involves managing and preserving other people's data, it would be embarrassing if I lost my own. So I've spent a lot of time over the years thinking about backup strategies for my personal files. This isn't a recommendation or a guide — just a description of what I do and why, in case it's useful to anyone thinking through similar questions.
The conventional wisdom in backup planning is the 3-2-1 rule: keep at least three copies of your data, on at least two different types of media, with at least one copy stored off-site. It's a solid framework. In practice, most people don't follow it, and I'll admit my own implementation is imperfect. But I try to stay close to the spirit of the rule even when the specifics vary.
My primary working machine is a desktop computer with a single internal solid state drive. This is where active project files, documents, email, and everything else lives day to day. The SSD is reasonably reliable — certainly more so than the spinning hard drives I used for years before switching — but any single drive can fail without warning. I've had it happen twice in my life, once with no backup at all (a painful lesson in my twenties), and once with a backup that was several weeks old (less painful, but I still lost some work). Those experiences were formative.
The first layer of my backup setup is a local external hard drive connected to the desktop via USB. It's a standard portable drive, nothing special, and it runs an incremental backup every evening using built-in operating system tools. The backup software creates versioned snapshots, so I can recover not just the most recent version of a file but also previous versions going back several weeks. This is the backup I use most often, and the scenario it covers is the most common one: I accidentally delete or overwrite a file, or I want to revert a document to an earlier state. It handles that well. The limitation is obvious — if something happens to the physical location where both the desktop and the external drive sit (fire, theft, flood, a sufficiently bad power surge), both copies are gone.
The second layer is another external hard drive that I keep in a different location. I update it less frequently — roughly once a month, sometimes less often if I'm busy. The process is manual: I bring the drive home, run a synchronization job that mirrors the contents of my primary backup drive to the off-site drive, verify that the sync completed, and then return the drive to its other location. It's not elegant. The gap between updates means I could lose up to a month of work if the worst happened. But it does satisfy the "off-site" part of the 3-2-1 rule, and for the types of files I work with, a month-old copy is vastly better than no copy at all.
I've considered cloud backup services as an alternative or supplement to the off-site drive. The appeal is obvious: continuous, automatic backup to a professionally managed facility with no physical transport required. I've tried a couple of these services. They work well enough for most files. My hesitation is partly about bandwidth — I work with some large media files, and the initial upload took an uncomfortably long time — and partly about the ongoing dependency on a third-party service. Cloud providers can change terms, raise prices, or go out of business. For now, I use a cloud sync service for a subset of my most critical documents, but I don't rely on it as my primary off-site solution.
One thing I've become more disciplined about over the years is verifying backups. A backup that you never test is a backup that might not work when you need it. I periodically pick a random selection of files from my backup drives and confirm that they open correctly and match the originals. I also occasionally do a test restore of a folder to make sure the backup software's restore process works as expected. This takes a small amount of time and provides a disproportionate amount of peace of mind. I've heard too many stories of people discovering that their backup was corrupted or incomplete only at the moment they needed it.
Drive rotation is something I've experimented with but haven't fully committed to. The idea is that you cycle through a set of three or more drives, retiring the oldest periodically and introducing a new one. This protects against the gradual degradation that all storage media experience over time. I currently have two drives in active use and one older drive in storage as a last-resort archive. I should probably be more systematic about this, but the cost adds up and I tend to keep using equipment until it shows signs of trouble.
File organization affects backup efficiency in ways I didn't appreciate early on. When your files are scattered across many locations — the desktop, downloads, various project directories — it's easy to miss something when configuring a backup job. I've moved toward keeping almost everything in a single directory tree, organized by project and year. This makes it straightforward to point the backup software at one top-level folder and know that everything important is included. It also makes the monthly sync to the off-site drive simpler.
I keep a plain text file at the root of my main data folder that lists what's in the backup, when it was last updated, and any notes about the current state of things. This sounds excessive, but it's saved me confusion more than once. When you come back to a backup drive after several months, it's surprisingly easy to forget what's on it, which version of your files it represents, and whether it was the drive you updated in March or the one you updated in June. A simple log file eliminates that ambiguity.
For particularly important files — tax documents, contracts, a few family photographs — I keep additional copies on a USB flash drive stored separately. Flash drives are not ideal for long-term archival storage, but as one more copy in a layered system, they serve a purpose.
None of this is sophisticated. There's no custom scripting, no enterprise-grade solution, no RAID array in a closet (though I've thought about it). It's just a set of habits built up over years of working with data and knowing what it feels like to lose some. The system works well enough for one person with a moderate amount of files and a low tolerance for data loss. If my needs were larger or my data more critical, I'd invest in something more robust. But for now, a couple of hard drives, a simple routine, and occasional verification keep me reasonably comfortable.
The main thing I'd say to anyone thinking about backups is: just start. An imperfect backup that exists is infinitely better than a perfect backup plan that you never get around to implementing. Copy your important files to an external drive today. You can optimize later.
— K. Vadala