generated html version of blog/optical_disks_for_data_archival.txtHOME

As you might know, I use optical disks routinely for archiving/backing up my
data. I've done this for 15 years, and it's been a success. But why use disks?
Well, This will be a (long) ramble on why I use optical disks as a backup 
system instead of other options.

Table of Contents
{A} Threat Model
{B} Backup Methods
{C} Optical Media
{D} More Rambling

{A} Threat Model

Why does one need to backup data? To some, this may seem like a silly question.
Of *course* one ought to have backups, to prevent data loss. But how does one
typically lose data? The most common one is not malicious, but simple 
misfortune. Computer hard disks are still the cheapest densest way to store 
data; at time of writing, a 2 terabyte magnetic hard disk can be had for less 
than 100 dollars, including an external plasitc box that connects by USB.

But magnetic hard disks can fail {A1} in ways such that the data is not 
recoverable except with the use of a cleanroom and extremely specialized 
equipment to repair their internals. They contain their own sophisticated 
recovery firmware to try to prevent this from happening, but nonetheless hard 
disk failures are still a threat.

The second threat {A2} I worry about is accidental deletion. When I want more 
space to store more stuff, it is common for me to look on my hard drive for 
unneeded files and just delete them. Deleting files for space on your hard disk 
requires you to not only delete the files, but also "empty the trash" as the 
metaphor goes. Once the space you've cleared by deleting the files has been 
overwritten by new files, the old files are gone.

A third threat, {A3} is malware. Various kinds of malware can make your files 
unusable or encrypt them and extort you. This form of malware is called 
"ransomware", and in recent years I have become more aware of high-profile 
incidents where companies were forced to pay criminals in cryptocurrency.

But even normal malware can infect files and scramble them. Some types of music
related software, made by supposedly non-criminal companies, has been known
to replace mp3 files with encrypted ones or different recordings downloaded
from their own web services. While I am adequately distrustful of these 
companies and do not use their music software, it is entirely possible for me 
to be tricked into using it or have it automatically installed.

{B} Backup Methods

So, given that I want to backup data, where should I store it, and what should
I store it on? The simplest one to implement at this time is probably an 
external drive. Flash drives are not fragile, and can store a pretty large 
amount of data, so they might be a good option. However, there are concerns 
about the longevity of data stored on one that has been left unpowered.
External hard disks have the same types of mechanical failures that are known 
to occur with internal hard disks.

Nevertheless, if you have these things lying around, storing backups on them is
an ok plan, so long as you continually update them. This protects pretty well
against threats {A1} and {A2}, but has an hitch when it comes to {A3}.
Malware can just as easily infect a USB-connected disk as it can an internal
drive. And in fact, many modern malware is designed specifically to be spread
from computer to computer via USB flash drives.

Malware cannot infect a USB flash drive if you fill it up and then put it away 
in a safe place. This precludes updating it constantly however. Cloud backup is 
a common solution and I do use it. However it has the issue that you need to 
continually pay for it, and hope that the company you store it with does not 
fail and take your data with it.

In theory, magnetic tape could be a possible backup medium-- however at time
of writing the cost of magnetic tape drives is prohibitive for the consumer.

{C} Optical Media

The solution I've used for a long time, is the writable optical disk, starting
with the CD, moving to the DVD, and most recently the Blu-Ray system. 
Originally, my chief motivation for this was cost. A DVD holds 4.3 GiB and 
costs about 30 canadian cents when bought in bulk today, thus about 7 cents
per gigabyte. At time of writing, an 128 GB flash thumb drive costs 20 canadian
dollars, or 15 cents per gigabyte. And flash storage is cheaper today than
ever before. However, compare Blu-Ray. A stack of 50 25GB BD-RE's can be had 
for around 50 dollars. This is 4 cents per GiB.

However, cost is not the only factor. Longevity is a concern for me. Flash 
drives are supposed to lose their data if left in a box after around 10 years,
although I have not tested this. Optical disks on the other hand have stated
longevity of 25 years, and I have cheap CD-R's that were burned in the 90's to
attest to this, which still work. I also have DVD-RAM disks manufactured in
2004 which continue to be both readable and writable over their entire capacity
in testing.

I suspect 25 years is not a limit but rather an conservative estimate. Magnetic
tape has been said to last 50 years, or even 100 years. Some types of optical
media have manufacturer-stated longevity of 100 years.

{D} More Rambling

As optical media is becoming less popular product, a small issue arises-
The Windows Explorer has become bad at writing reliably to optical drives.
I recommend formatting the disks as UDF version 2.5 and writing using 
alternative file manager such as Total Commander; or using an specialized
disk burning program. Linux of course has no issue whatsoever writing files 
to optical disks, without needing special software.

Storing optical disks is best kept in a cool dry place. I recommend paper 
or cardboard sleeves without plastic windows, as these windows have been known
to stick to the surface of the disks. I label my rewritable disks with washable 
crayons or pencil crayons, since then if I reuse a disk to store something 
different, I can wipe off the old label and write a new label.

The solution in the past typically for disk labeling was the sharpie(TM) 
permanent marker. If you find yourself needing to remove these marks, I 
recommend soaking in isopropyl alcohol and then wiping with a cotton ball.

One of the major things I am backing up is videos from my favourite youtubers,
because I do not trust Google's current management to be kind to them or keep
their videos available. I also backup my Steam games, because Steam could 
potentially take games offline (although, currently I trust Valve far more than 
Google in this regard.)

Optical disks are said to be prone to scratches. In my experience this is only
true if you leave them on your desk out of a sleeve. One tip- the most 
vulnerable part of a CD is actually the label side, since the data is written
just under the label. Optical disks stored well can be entirely scratch free
even after decades of use. If you do have disks that are scratched, simply rip 
them to your hard disk and burn them on a new fresh disk. The only exceptions
that tend to cause problems are game console disks. For those, emulation is
the best solution.

Currently, the newest type of disk is the BD-XL format, which corresponds to 
the commercial Ultra-HD blu-ray format for movies. The BD-RE XL stores 33 GB per 
layer with 3 layers, thus 93 GiB. There are also 4 layer non-rewritable disks 
with a capacity of 128 GB or 119 GiB. These are not compatible with older 
blu-ray burners.

There have been continual rumours since the early 2000's of so-called 
holographic or 3D disks with terabytes or even petabytes of storage per disk.
This has never materialized so far, but if it ever does, one might see it 
replace magnetic hard drives.

Oren Watson 
2023 day 62