The Myth of the Immortal Server: Why Digital Preservation Is Not a Backup Plan
There’s a comforting, pervasive idea in our digital age: if something is important, we back it up. We copy files to external drives, sync them to the cloud, and trust in the redundancy of distributed systems. This logic has been casually extended to our collective digital heritage. The received wisdom goes something like this: if a website, a dataset, or a digital publication is valuable, someone, somewhere, will have saved it. It’s out there, backed up, waiting to be rediscovered. This is a dangerous and profound misconception.
Digital preservation is not backup. A backup is a short-term, operational strategy for disaster recovery. Its goal is continuity, ensuring that a business or service can resume after a hard drive fails or a server room floods. Preservation, by contrast, is a long-term, strategic commitment to maintaining not just the bits of a file, but its meaning, context, and usability across decades or even centuries. A backup ensures a file exists; preservation ensures it can be understood.
Consider a simple example: a backup of a government report from 1995 might perfectly preserve a .pdf file. But without also preserving the specific version of Adobe Acrobat Reader needed to open it, the emulation software to run that obsolete program, and the documentation to explain its peculiar formatting, the file is just a sealed digital tomb. The bits are safe, but the information is entombed. A backup cares for the container; preservation must care for the key.
This myth fosters a dangerous complacency. It allows institutions and individuals to assume the work is done once something is copied to a second location. It ignores the relentless march of technological obsolescence, the silent decay of file formats, and the subtle erosion of contextual knowledge that gives data its significance. The most critical work of preservation isn’t the initial act of saving; it’s the ongoing, unglamorous labor of maintenance, migration, and metadata enrichment. It’s the work of ensuring that a JPEG doesn’t become an abstract collection of pixels whose purpose is forgotten, or that a complex relational database doesn’t become an indecipherable digital artifact, its structure lost without the proprietary software that gave it life.
By conflating backup with preservation, we risk creating vast digital graveyards—petabytes of perfectly stored, utterly useless data. We amass the containers but lose the keys en masse. True digital preservation requires a shift in mindset, from a one-time action to a perpetual process. It demands acknowledging that saving something is only the very first step in a never-ending relay race against time. The myth of the immortal server tells us our work is done when we hit ‘copy.’ The reality is that the work has only just begun.
Notes & further reading
A few pages I came back to while writing this: