Some odd number of years ago, I read a Reddit post12 that greatly influenced the way I think about digital preservation and the data we create and use on a daily basis. For many in photography communities, it’s not uncommon to create hundreds of gigabytes of raw image data over the course of several years. For working photographers to lose photos, this can literally be their livelihood — maybe there’s photos of a client’s wedding they just lost, or a unique editing configuration in Lightroom. In many cases they just have sentimental value — maybe they document stages of life or a special trip.
I’ve had two hard drive incidents happen before. One was in high school — where a faulty SD card led to my (rather unreliable old Android 4.4) phone to wipe itself, losing 2FA keys and a good amount of data. The other was, at the time of writing, fairly recent. It happened in a way that is so ridiculous to describe but here we are: I finished with my work one day, shut my laptop screen to go to the washroom, came back and suddenly found myself with no SSD detected.
In both cases, what saved me was a robust (at least comparably speaking) cloud back-up system. To credit my old phone, it came with Xiaomi Cloud, provided by the manufacturer. For the second time, I had gotten so good at backing-up much of my work and notes to OneDrive and GitHub or simply never working offline that in terms of data I lost very little.
When we think about large data systems, there’s a few big things we want to consider. What happens when we lose data? How can we maintain the security and integrity of data? This goes for our own personal systems too. Over the course of our lifetimes, we generate huge amounts of photos and work that can’t be lost.
There is no single storage device that is infallible. In fact, using USB drives and SSDs and any sort of local storage is extremely failure prone (part of this is due to the nature of the hardware). What happens when that one USB drive is lost or damaged or destroyed?
So we discuss back-ups. When (not if) our data storage systems fail, we should have something we can get back. Cloud storage (Google Drive, OneDrive, Backblaze), regular pushes to a Git repository, physical back-ups off-site (at a relative’s house). Robust and resilient systems spanning multiple points mean that if one thing fails the rest doesn’t.
TL;DR — It does not suffice to have one local version of anything. Back-up, preferably to the cloud.
Written 6 November 2023
Keywords:
- Data
- Photography
- Digital preservation
- Multi-factor authentication
- Cloud storage
- Git
- Software systems
Footnotes
-
It was this one specifically: Hard drive failures have no mercy, sense of time, or value of the media on it ↩
-
This one is an equally applicable read: Look At Your Computer/Hard Drive. What If It Caught Fire Right Now? ↩