Long term Data storage
data rot is a massive concern no one is concerned about
I know a lot of people are not aware that their digital information can be gone in an instant, at least not until it happens for the first time.
You save your data on some external hard drive, thinking it’s safe, but a decade later, half of it is unreadable. You take out your CD where you saved your favorite moments and it decomposed.
We forget just how fragile our digital world is. This was pushed from our minds with more prevalent cloud storage, where we offloaded the care of our data to massive data centers, but that comes at other upfront and hidden costs. As the “old” saying goes, if it’s free or cheap, then the product is you.
CDs and DVDs are thought by most people as archival media, but that’s far from reality. If it’s left open and exposed to elements, optical discs deteriorate rapidly. It can warp from heat in minutes, and UV rays can corrupt data in days.
The reflective layer used to read data can oxidize with age, making the disk impossible to read. And, moisture can enter and delaminate the layers. It’s not a matter of if but when those things occur, regardless of how protective you are in storing them outside of specialized salt mine storage facilities.
Mechanical Hard drives are also susceptible to data rot, among other things. Data is written with magnetic fields that can weaken over time, causing random bit flips that corrupt files, making them unreadable. It takes time, over 10 years, but that’s still fast.
While it is possible to recover such files, some information will be lost. For example, this could result in images loading halfway or misaligning. Then, mechanical parts of the drive can wear out, lubrication can degrade, or simple physical damage can happen when you accidentally drop it. There are different ways to mitigate those situations, but in the end, it’s a mechanical drive that rotates thousands of times a minute with a read/write head hovering above it, reading magnetic data; it’s an error waiting to happen.
Solid State Drives are super fast, but are super unreliable at unpowered data storage. SSD data retention without power is limited, typically 1 year for consumer drives at 30°C. They use flash memory, which can lose data over time due to charge leakage in the memory cells, especially if not powered on regularly. If you store it in a cold, dry place and power it now and again, it should be fine for a decade or so.
Be it magnetic or etched in plastic, our data is not safe, and we can’t expect to pass it down to our grandchildren like they did with books and photo albums. Paper, for all its fragility, is an excellent store of data. We still have books that are thousands of years old and more. There are archival grade cotton papers (100 % cotton fiber) that can last > 1,000 years in optimal conditions.
There is a good writeup on legendary blog Coding Horror by Jeff Atwood discussing saving data on paper with info from PaperBack, a software to do just that, that with a good laser printer with the 600 dpi resolution, you can save up to 500,000 bytes, or about 500kb, of uncompressed data on a single sheet. Another software that can store data on paper is Twibright Optar. A more detailed writeup about different encoding approaches is by Martin Monperrus How to store data on paper?.
I’m currently working on my own way of storing data on paper, for black and white and for colored printers. Colored printers can store more data, but decoding them is a bit tricky. That project is in progress, and I will update with a new post on the progress. If you want to follow along, here is the GitHub repo link.
Another good medium for storage is glass or crystals. Microsoft has its Silica project that laser etches data to “glass” like medium. The issue is, it’s not commercially available, and even if it were, I doubt it would be within the reach of most users.
While doing research for this post, I stumbled upon mDisc. Launched in 2009 by Millenniata, mDisc promised a revolution in archival storage. Unlike standard DVDs or Blu-rays that degrade in 10–20 years due to dye breakdown, mDisc’s data layer withstands light, heat, and humidity. Tests, like the U.S. Department of Defense’s 2009 study, showed it outlasting competitors, with claims of 100–1,000 years of data retention. It’s available in 4.7GB DVDs and 25GB/50GB Blu-ray formats, with Verbatim and Ritek as key manufacturers today.
Millenniata hit a wall in 2016, filing for bankruptcy after defaulting on convertible debt. Debt holders took over, forming Yours.co to continue mDisc sales, but that venture fizzled out by the early 2020s, and its Kickstarter was abandoned, its website a ghost town. The original mdisc.com lingers, but it’s a shadow of its former self, with broken links and no updates. Despite this, mDisc lives on through licensed partners like Verbatim, which still produces co-branded discs.
In the end, it boils down to this: our digital treasures, photos, documents, or code aren’t immortal just because they live on drives, discs, or especially in “the cloud.” Every medium has a breaking point: magnetic fields fade, dye layers decay, and flash cells leak their charge. If you truly want your great-grandchildren to glimpse your memories, you’ve got to think beyond convenience and plan for centuries, not years.
M-Disc still stands out as the only consumer-friendly optical format engineered for the long haul, and glass based solutions hold tremendous promise. But still the safest backup is the one you’re willing to revisit, verify, and migrate every decade or so.