Photos, videos, banking, life history, assignments etc. It’s a wildly misunderstood process which shouldn't be over complicated. It can be summed up pretty nicely with the 3, 2, 1 strategy – 3 copies, 2 different media and (at least) 1 offsite. It doesn't need to be that hard and at th end I give a couple of 'pro' tips to give an extra layer pf protection…
Let’s start at the beginning, you need backups because if your hard drive (or external storage) fails you’ll potentially loose everything. Let’s work through a couple of ways to protect yourself from computer karma.
First up, a backup can be in many forms - the most basic of which is a simple external drive attached to your computer where you can keep a COPY of the files on your computer. For this to be a backup it NEEDS to be a COPY of the data. If you're simple storing unique files on an external drive it's just as risky as keeping those unique files on your computer's disk.
Archived (or staged backups) do more than a basic copy of files. Typically they do an initial full backup and then will incrementally copy file changes. The most notable of these is Apple's Time Machine. You essentially plug in a drive, tell Time Machine to use this disk for backups and it will do the rest. The benefit of this type of backup is the ability ti go back to previous versions of the file. If a file gets corrupted you might be able to go back to a prior version without the corruption.
To add some additional safety (and complexity) you can go the next leve and use RAID.
RAID stands for redundant array of inexpensive(independent) disks - yes, there are actually two definition but let's not nit pick. RAID is like black magic to most people but in very basic terms it’s a way to save your data over multiple disks/volumes to protect against the failure of one of those disks. Here’s a brief rundown:
RAID 0 (striping) – if you’re an average user, don’t use RAID 0. RAID 0 is designed for maximum performance but is a ‘total loss’ scenario in the event of ANY disk failure in the array. It writes unique pieces of your data to different drives but if one fails you lose everything. Think of it as half your data is on each disk and if one fails you can no longer retrieve the complete file.Once again DO NOT use this for normal use. There are exceptions where RAID 0 is good but it's lousy for backups. RAID 0 requires two drives which combine to the size you need. If you want 4 TB of storage you’ll need two 2 TB drives.
RAID 1 (Mirroring) – RAID 1 is the most likely option for the average user. It writes complete files to two separate disks which form part of a single, mounted volume. In other words, there are two disks combined into a RAID arrray but on your computer you will just see a single volume. If one disk fails, the other disk will keep working as though nothing has happened. Generally the RAID will notify you of a problem which you can address by fixing the failed drive. When this is done, the drive will ‘rebuild’ itself to a fully functional state. The best feature of RAID 1 is you can continue to work normally if one disk fails.
It’s important to know RAID 1 mirrors are real time – when you save a file it is automatically written to both drives giving very good protection against a single disk failure. On the downside it also means if you have a corrupt file it will be corrupt on both drives. RAID 1 is slightly slower than normal disk operations because it needs to write to two places instead of one. This is minimised by the use of caches and most people will struggle to tell the difference. RAID 1 requires two disks of the size you need. If you need to have 4 TB of storage you’ll need two 4TB drives.
RAID 5 (distributed parity) – RAID 5 is great if you don’t mind spending a bit more money. RAID 5 typically uses 4 or 5 drives over which the data is distributed. A parity bit is shared so in the event of a drive failure, the RAID array can keep operating as it knows what the missing piece is. Think of it like a Sudoku puzzle where you know 6 of the numbers on a line then you must know the seventh.
One of the beauties of RAID 5 is the ability to have a ‘hot’ backup. If you have a 5 disk array, 4 of them can be combined into the main RAID and the fifth can be used as a hot spare. If a drive fails the RAID will automatically implement the spare drive and update it to be current. This means you can happily keep working and replace the drive with no down time.
RAID 5 requires more physical drives. Due to the configuration requirements you lose 25% of the total capacity of the drives installed. To get 4 TB of RAID 5 you’ll need (roughly) 4x 1.5 TB drives. If you have a 5 disk enclosure and want a hot spare you’ll need 5 drives. RAID 5 is great but probably overkill for the average user.
If you want a RAID device, try OWC – https://eshop.macsales.com/shop/owc-mercury-elite-pro-dock (Note these come configured as RAID 0 so the capacity can be misleading, it will appear double what a RAID 1 would be)
Whichever RAID you choose, make sure to buy something with an appropriate connection for your machine and preferably one with some future to it.
TL:DR Use RAID 1 mirroring. You get excellent backup of your data at a reasonable cost.
Whilst RAID is great, it’s not really a backup, rather it’s a redundancy. It can be considered ONE of the two copies you need to comply with the 3, 2, 1 rule. Your second copy can be any other form of disk. A simple attached USB drive will suffice as long as you copy data to it regularly. Ideally, you’ll store this away from your computer – in a detached shed, in a fireproof safe etc. Once you have this you are two thirds the way to being protected.
The final piece of the puzzle if off site backup. As much as we don’t want to consider it, houses can burn down or lightning strikes can nuke your data drives. This is where offsite backups come in. The media doesn’t matter too much as long as it can store everything you need. Make a copy of your files and take it to a friend or relative’s place or maybe to your office at work. You have to remember to keep this updated whenever you do major updates to your files (like dumping an SD card’s worth of your family’s round the world holiday photos).
The ‘perfect’ way to do this is to have two offsite drives. One stays offsite and the others stays with your computer and gets backed up. Next time you go to your offsite location take the disk from your computer and swap it with the one from offsite.
Back in the days when I was a photographer it would cost around $20 for a roll of film – 36 exposures. It was then $10 to get it processed. I often shot 30 or 40 rolls on a trip so could easily drop $1000. Now consider what an SD card costs – I can get a 32GB card for $13. That 32GB will get me (literally) thousands of images. Here’s my tip – NEVER reuse an SD card. When it’s full, stick it in an envelope, label it and put it aside somewhere cool and dry. This is your ‘extra’ backup and also serves as a TOTAL backup for those times when you accidentally delete something on your computer which you should have kept.
Pro Tip two
OK, this is not so much a ‘pro tip’ but is a quick way to be extra protected. I shoot RAW and process everything through Capture One. I rate and tag everything and when I’m done, I get all the 5 star rated images and export a reasonably high resolution jpg file of each. Being on a Mac, I then dump them into Apple Photos. Now I don’t like Photos BUT it can automatically backup to iCloud (if you have sufficient storage) which provides a decent offsite backup. It has the added benefit of making all your best images available on all your devices – Macs, iPhone, iPad etc. If you set up family sharing, all your family get to see them. At a pinch, those good quality jpgs can also serve well for prints or albums etc.